use strict;
use warnings;
-our $VERSION = "0.20";
+our $VERSION = "0.31";
our @ISA = qw(Exporter);
our @EXPORT_OK = ('regmust',
qw(is_regexp regexp_pattern
s => 1 << ($PMMOD_SHIFT + 1),
i => 1 << ($PMMOD_SHIFT + 2),
x => 1 << ($PMMOD_SHIFT + 3),
- p => 1 << ($PMMOD_SHIFT + 4),
+ n => 1 << ($PMMOD_SHIFT + 5),
+ p => 1 << ($PMMOD_SHIFT + 6),
+ strict => 1 << ($PMMOD_SHIFT + 10),
# special cases:
d => 0,
l => 1,
TRIEC => 0x000004,
DUMP => 0x000008,
FLAGS => 0x000010,
+ TEST => 0x000020,
EXECUTE => 0x00FF00,
INTUIT => 0x000100,
sub bits {
my $on = shift;
my $bits = 0;
+ my $turning_all_off = ! @_ && ! $on;
+ my %seen; # Has flag already been seen?
+ if ($turning_all_off) {
+
+ # Pretend were called with certain parameters, which are best dealt
+ # with XXX
+ push @_, keys %bitmask; # taint and eval
+ push @_, 'strict';
+ }
+
+ # Process each subpragma parameter
ARG:
foreach my $idx (0..$#_){
my $s=$_[$idx];
} elsif ($EXPORT_OK{$s}) {
require Exporter;
re->export_to_level(2, 're', $s);
+ } elsif ($s eq 'strict') {
+ if ($on) {
+ $^H{reflags} |= $reflags{$s};
+ warnings::warnif('experimental::re_strict',
+ "\"use re 'strict'\" is experimental");
+
+ # Turn on warnings if not already done.
+ if (! warnings::enabled('regexp')) {
+ require warnings;
+ warnings->import('regexp');
+ $^H{re_strict} = 1;
+ }
+ }
+ else {
+ $^H{reflags} &= ~$reflags{$s} if $^H{reflags};
+
+ # Turn off warnings if we turned them on.
+ warnings->unimport('regexp') if $^H{re_strict};
+ }
+ if ($^H{reflags}) {
+ $^H |= $flags_hint;
+ }
+ else {
+ $^H &= ~$flags_hint;
+ }
} elsif ($s =~ s/^\///) {
my $reflags = $^H{reflags} || 0;
my $seen_charset;
while ($s =~ m/( . )/gx) {
- $_ = $1;
+ local $_ = $1;
if (/[adul]/) {
# The 'a' may be repeated; hide this from the rest of the
# code by counting and getting rid of all of them, then
}
else {
delete $^H{reflags_charset}
- if defined $^H{reflags_charset}
- && $^H{reflags_charset} == $reflags{$_};
+ if defined $^H{reflags_charset}
+ && $^H{reflags_charset} == $reflags{$_};
}
} elsif (exists $reflags{$_}) {
- $on
+ $seen{$_}++;
+ $on
? $reflags |= $reflags{$_}
: ($reflags &= ~$reflags{$_});
} else {
}
}
($^H{reflags} = $reflags or defined $^H{reflags_charset})
- ? $^H |= $flags_hint
- : ($^H &= ~$flags_hint);
+ ? $^H |= $flags_hint
+ : ($^H &= ~$flags_hint);
} else {
require Carp;
Carp::carp("Unknown \"re\" subpragma '$s' (known ones are: ",
")");
}
}
+ if (exists $seen{'x'} && $seen{'x'} > 1
+ && (warnings::enabled("deprecated")
+ || warnings::enabled("regexp")))
+ {
+ my $message = "Having more than one /x regexp modifier is deprecated";
+ if (warnings::enabled("deprecated")) {
+ warnings::warn("deprecated", $message);
+ }
+ else {
+ warnings::warn("regexp", $message);
+ }
+ }
+
+ if ($turning_all_off) {
+ _load_unload(0);
+ $^H{reflags} = 0;
+ $^H{reflags_charset} = 0;
+ $^H &= ~$flags_hint;
+ }
+
$bits;
}
$pat = '(?{ $foo = 1 })';
use re 'eval';
- /foo${pat}bar/; # won't fail (when not under -T switch)
+ /foo${pat}bar/; # won't fail (when not under -T
+ # switch)
{
no re 'taint'; # the default
($x) = ($^X =~ /^(.*)$/s); # $x is not tainted here
no re 'eval'; # the default
- /foo${pat}bar/; # disallowed (with or without -T switch)
+ /foo${pat}bar/; # disallowed (with or without -T
+ # switch)
}
+ use re 'strict'; # Raise warnings for more conditions
+
use re '/ix';
"FOO" =~ / foo /; # /ix implied
no re '/x';
"FOO" =~ /foo/; # just /i implied
use re 'debug'; # output debugging info during
- /^(.*)$/s; # compile and run time
+ /^(.*)$/s; # compile and run time
- use re 'debugcolor'; # same as 'debug', but with colored output
+ use re 'debugcolor'; # same as 'debug', but with colored
+ # output
...
- use re qw(Debug All); # Finer tuned debugging options.
- use re qw(Debug More);
- no re qw(Debug ALL); # Turn of all re debugging in this scope
+ use re qw(Debug All); # Same as "use re 'debug'", but you
+ # can use "Debug" with things other
+ # than 'All'
+ use re qw(Debug More); # 'All' plus output more details
+ no re qw(Debug ALL); # Turn on (almost) all re debugging
+ # in this scope
use re qw(is_regexp regexp_pattern); # import utility functions
my ($pat,$mods)=regexp_pattern(qr/foo/i);
if (is_regexp($obj)) {
print "Got regexp: ",
- scalar regexp_pattern($obj); # just as perl would stringify it
- } # but no hassle with blessed re's.
+ scalar regexp_pattern($obj); # just as perl would stringify
+ } # it but no hassle with blessed
+ # re's.
(We use $^X in these examples because it's tainted by default.)
I<is> allowed if $pat is a precompiled regular expression, even
if $pat contains C<(?{ ... })> assertions or C<(??{ ... })> subexpressions.
+=head2 'strict' mode
+
+When C<use re 'strict'> is in effect, stricter checks are applied than
+otherwise when compiling regular expressions patterns. These may cause more
+warnings to be raised than otherwise, and more things to be fatal instead of
+just warnings. The purpose of this is to find and report at compile time some
+things, which may be legal, but have a reasonable possibility of not being the
+programmer's actual intent. This automatically turns on the C<"regexp">
+warnings category (if not already on) within its scope.
+
+As an example of something that is caught under C<"strict'> but not otherwise
+is the pattern
+
+ qr/\xABC/
+
+The C<"\x"> construct without curly braces should be followed by exactly two
+hex digits; this one is followed by three. This currently evaluates as
+equivalent to
+
+ qr/\x{AB}C/
+
+that is, the character whose code point value is C<0xAB>, followed by the
+letter C<C>. But since C<C> is a a hex digit, there is a reasonable chance
+that the intent was
+
+ qr/\x{ABC}/
+
+that is the single character at C<0xABC>. Under C<'strict'> it is an error to
+not follow C<\x> with exactly two hex digits. When not under C<'strict'> a
+warning is generated if there is only one hex digit, and no warning is raised
+if there are more than two.
+
+It is expected that what exactly C<'strict'> does will evolve over time as we
+gain experience with it. This means that programs that compile under it in
+today's Perl may not compile, or may have more or fewer warnings, in future
+Perls. There is no backwards compatibility promises with regards to it. For
+this reason, using it will raise a C<experimental::re_strict> class warning,
+unless that category is turned off.
+
+Note that if a pattern compiled within C<'strict'> is recompiled, say by
+interpolating into another pattern, outside of C<'strict'>, it is not checked
+again for strictness. This is because if it works under strict it must work
+under non-strict.
+
=head2 '/flags' mode
When C<use re '/flags'> is specified, the given flags are automatically
Dump the final program out after it is compiled and optimised.
+=item FLAGS
+
+Dump the flags associated with the program
+
+=item TEST
+
+Print output intended for testing the internals of the compile process
+
=back
=item Execute related options
=item INTUIT
-Enable debugging of start point optimisations.
+Enable debugging of start-point optimisations.
=back
or disabling this option automatically does the same for debugging
states as well. This output from this can be quite large.
+=item GPOS
+
+Enable debugging of the \G modifier.
+
=item OPTIMISEM
-Enable enhanced optimisation debugging and start point optimisations.
+Enable enhanced optimisation debugging and start-point optimisations.
Probably not useful except when debugging the regexp engine itself.
=item OFFSETS
Almost definitely only useful to people hacking
on the offsets part of the debug engine.
+
=back
=item Other useful flags
=item ALL
-Enable all options at once except OFFSETS, OFFSETSDBG and BUFFERS
+Enable all options at once except OFFSETS, OFFSETSDBG and BUFFERS.
+(To get every single option without exception, use both ALL and EXTRA.)
=item All
=item More
-Enable TRIEM and all execute compile and execute options.
+Enable the options enabled by "All", plus STATE, TRIEC, and TRIEM.
=back
=back
As of 5.9.5 the directive C<use re 'debug'> and its equivalents are
-lexically scoped, as the other directives are. However they have both
+lexically scoped, as are the other directives. However they have both
compile-time and run-time effects.
=head2 Exportable Functions
Because the C<here> is before the C<.*> in the pattern, its position
can be determined exactly. That's not true, however, for the C<there>;
it could appear at any point after where the anchored string appeared.
-Perl uses both for its optimisations, prefering the longer, or, if they are
+Perl uses both for its optimisations, preferring the longer, or, if they are
equal, the floating.
B<NOTE:> This may not necessarily be the definitive longest anchored and