But some description
Some description that is not output
--- selector
p *:contains("description")
--- expected
two level deep description
But some description
===
--- input
No (no preceding sibling)
A header
Yes
Some description
Another two level deep description
Foobar
Barred
Some more description
No (child, not sibling)
Some description that is not output
--- selector
p > *:contains("description")
--- expected
Some more description
===
--- input
No (no preceding sibling)
A header
Yes
Some description
Another two level deep description
Foobar
Barred
No (child, not sibling)
Some more description
--- selector
*:contains("description")
--- expected
Some description
two level deep description
Some more description
===
--- input
Some description
Another two level deep description
Some more description
--- selector
:empty
--- expected
===
--- input
here
--- selector
div *:not(p) em
--- expected
here
===
--- input
here
--- selector
//div/*[not(self::p)]/em
--- expected
here
===
--- input
here
there
--- selector
div em:only-child
--- expected
everywhere
===
--- input
here
there
everywhere
elsewhere
nowhere
--- selector
div em:nth-child(2n)
--- expected
there
elsewhere
===
--- input
here
there
everywhere
elsewhere
nowhere
--- selector
div em:nth-child(2n+1)
--- expected
here
everywhere
nowhere
===
--- input
here
there
everywhere
elsewhere
nowhere
anywhere
--- selector
div em:nth-last-child(3n)
--- expected
here
elsewhere
===
--- input
anywhere
here
there
everywhere
elsewhere
nowhere
--- selector
div em:nth-last-child(2n+1)
--- expected
here
everywhere
nowhere
HTML-Selector-XPath-0.16/t/03_xpath_root.t 0000644 0001750 0001750 00000001407 12201461127 017467 0 ustar corion corion use strict;
use Test::Base;
use HTML::Selector::XPath qw(selector_to_xpath);
plan tests => 1 * blocks;
filters { selector => 'chomp', xpath => 'chomp' };
run {
my $block = shift;
is selector_to_xpath($block->selector, root=> '/R'), $block->xpath, $block->selector;
}
__END__
===
--- selector
*
--- xpath
/R/*
===
--- selector
E
--- xpath
/R/E
===
--- selector
E F
--- xpath
/R/E//F
===
--- selector
E > F
--- xpath
/R/E/F
===
--- selector
E + F
--- xpath
/R/E/following-sibling::*[1]/self::F
===
--- selector
E[foo]
--- xpath
/R/E[@foo]
===
--- selector
E[foo="warning"]
--- xpath
/R/E[@foo='warning']
===
--- selector
E#myid
--- xpath
/R/E[@id='myid']
===
--- selector
foo.bar, bar
--- xpath
/R/foo[contains(concat(' ', @class, ' '), ' bar ')] | /R/bar
HTML-Selector-XPath-0.16/t/01_xpath.t 0000644 0001750 0001750 00000021055 12201461127 016423 0 ustar corion corion use strict;
use Test::Base;
use HTML::Selector::XPath;
my @should_die = split /\n+/, <<'EOF';
[1a]
[-1a]
[--a]
[!a]
[ab!c]
[]
[x=1a]
[x=-1a]
[x=--a]
[x=!a]
[x=ab!c]
[x="]
[x="abc" "]
[x=abc z]
EOF
plan tests => 1 * blocks() + @should_die;
filters { selector => 'chomp', xpath => 'chomp' };
run {
my $block = shift;
my $selector = HTML::Selector::XPath->new($block->selector);
is $selector->to_xpath, $block->xpath, $block->selector;
};
for my $selector (@should_die) {
my $to_xpath = eval { HTML::Selector::XPath->new($selector)->to_xpath };
is($to_xpath, undef, "invalid selector should die: $selector");
}
__END__
===
--- selector
*
--- xpath
//*
===
--- selector
E
--- xpath
//E
===
--- selector
E F
--- xpath
//E//F
===
--- selector
E > F
--- xpath
//E/F
===
--- selector
p.pastoral.marine
--- xpath
//p[contains(concat(' ', @class, ' '), ' pastoral ')][contains(concat(' ', @class, ' '), ' marine ')]
===
--- selector
E:first-child
--- xpath
//E[count(preceding-sibling::*) = 0 and parent::*]
===
--- selector
F E:first-child
--- xpath
//F//E[count(preceding-sibling::*) = 0 and parent::*]
===
--- selector
F > E:first-child
--- xpath
//F/E[count(preceding-sibling::*) = 0 and parent::*]
===
--- selector
E:lang(c)
--- xpath
//E[@xml:lang='c' or starts-with(@xml:lang, 'c-')]
===
--- selector
E + F
--- xpath
//E/following-sibling::*[1]/self::F
===
--- selector
E + #bar
--- xpath
//E/following-sibling::*[1]/self::*[@id='bar']
===
--- selector
E + .bar
--- xpath
//E/following-sibling::*[1]/self::*[contains(concat(' ', @class, ' '), ' bar ')]
===
--- selector
E[foo]
--- xpath
//E[@foo]
===
--- selector
E[foo=warning]
--- xpath
//E[@foo='warning']
===
--- selector
E[foo="warning"]
--- xpath
//E[@foo='warning']
===
--- selector
E[foo~="warning"]
--- xpath
//E[contains(concat(' ', @foo, ' '), ' warning ')]
===
--- selector
E[foo~=warning]
--- xpath
//E[contains(concat(' ', @foo, ' '), ' warning ')]
===
--- selector
E[foo^="warning"]
--- xpath
//E[starts-with(@foo,'warning')]
===
--- selector
E[foo^=warning]
--- xpath
//E[starts-with(@foo,'warning')]
===
--- selector
E:not([foo^="warning"])
--- xpath
//E[not(starts-with(@foo,'warning'))]
===
--- selector
E:not([foo^=warning])
--- xpath
//E[not(starts-with(@foo,'warning'))]
===
--- selector
E[foo$="warning"]
--- xpath
//E[ends-with(@foo,'warning')]
===
--- selector
E[foo$=warning]
--- xpath
//E[ends-with(@foo,'warning')]
===
--- selector
E[lang|="en"]
--- xpath
//E[@lang='en' or starts-with(@lang, 'en-')]
===
--- selector
E[lang|=en]
--- xpath
//E[@lang='en' or starts-with(@lang, 'en-')]
===
--- selector
DIV.warning
--- xpath
//DIV[contains(concat(' ', @class, ' '), ' warning ')]
===
--- selector
E#myid
--- xpath
//E[@id='myid']
===
--- selector
foo.bar, bar
--- xpath
//foo[contains(concat(' ', @class, ' '), ' bar ')] | //bar
===
--- selector
E:nth-child(1)
--- xpath
//E[count(preceding-sibling::*) = 0 and parent::*]
===
--- selector
E:last-child
--- xpath
//E[count(following-sibling::*) = 0 and parent::*]
===
--- selector
F E:last-child
--- xpath
//F//E[count(following-sibling::*) = 0 and parent::*]
===
--- selector
F > E:last-child
--- xpath
//F/E[count(following-sibling::*) = 0 and parent::*]
===
--- selector
E[href*="bar"]
--- xpath
//E[contains(@href, 'bar')]
===
--- selector
E[href*=bar]
--- xpath
//E[contains(@href, 'bar')]
===
--- selector
E:not([href*="bar"])
--- xpath
//E[not(contains(@href, 'bar'))]
===
--- selector
E:not([href*=bar])
--- xpath
//E[not(contains(@href, 'bar'))]
===
--- selector
F > E:nth-of-type(3)
--- xpath
//F/E[3]
===
--- selector
E ~ F
--- xpath
//E/following-sibling::F
===
--- selector
E ~ F.foo
--- xpath
//E/following-sibling::F[contains(concat(' ', @class, ' '), ' foo ')]
===
--- selector
E:contains("Hello")
--- xpath
//E[text()[contains(string(.),"Hello")]]
===
--- selector
E:contains( "Hello" )
--- xpath
//E[text()[contains(string(.),"Hello")]]
===
--- selector
E:contains( "Hello" ).C
--- xpath
//E[text()[contains(string(.),"Hello")]][contains(concat(' ', @class, ' '), ' C ')]
===
--- selector
E:contains( "Hello" ) .C
--- xpath
//E[text()[contains(string(.),"Hello")]]//*[contains(concat(' ', @class, ' '), ' C ')]
===
--- selector
F, E:contains( "Hello" )
--- xpath
//F | //E[text()[contains(string(.),"Hello")]]
===
--- selector
E:contains( "Hello" ), F
--- xpath
//E[text()[contains(string(.),"Hello")]] | //F
===
--- selector
E ~ F
--- xpath
//E/following-sibling::F
===
--- selector
E ~ #bar
--- xpath
//E/following-sibling::*[@id='bar']
===
--- selector
E ~ .bar
--- xpath
//E/following-sibling::*[contains(concat(' ', @class, ' '), ' bar ')]
===
--- selector
E ~ *
--- xpath
//E/following-sibling::*
===
--- selector
.foo ~ E
--- xpath
//*[contains(concat(' ', @class, ' '), ' foo ')]/following-sibling::E
===
--- selector
.foo ~ *
--- xpath
//*[contains(concat(' ', @class, ' '), ' foo ')]/following-sibling::*
===
--- selector
.foo ~ .bar
--- xpath
//*[contains(concat(' ', @class, ' '), ' foo ')]/following-sibling::*[contains(concat(' ', @class, ' '), ' bar ')]
===
--- selector
> em
--- xpath
//*/em
===
--- selector
:first-child
--- xpath
//*[count(preceding-sibling::*) = 0 and parent::*]
===
--- selector
:last-child
--- xpath
//*[count(following-sibling::*) = 0 and parent::*]
===
--- selector
E.c:first-child
--- xpath
//E[contains(concat(' ', @class, ' '), ' c ')][count(preceding-sibling::*) = 0 and parent::*]
===
--- selector
E:first-child.c
--- xpath
//E[count(preceding-sibling::*) = 0 and parent::*][contains(concat(' ', @class, ' '), ' c ')]
===
--- selector
E#i:first-child
--- xpath
//E[@id='i'][count(preceding-sibling::*) = 0 and parent::*]
===
--- selector
E:first-child#i
--- xpath
//E[count(preceding-sibling::*) = 0 and parent::*][@id='i']
===
--- selector
:lang(c)
--- xpath
//*[@xml:lang='c' or starts-with(@xml:lang, 'c-')]
===
--- selector
:lang(c)#i
--- xpath
//*[@xml:lang='c' or starts-with(@xml:lang, 'c-')][@id='i']
===
--- selector
#i:lang(c)
--- xpath
//*[@id='i'][@xml:lang='c' or starts-with(@xml:lang, 'c-')]
===
--- selector
*:lang(c)#i
--- xpath
//*[@xml:lang='c' or starts-with(@xml:lang, 'c-')][@id='i']
===
--- selector
E:lang(c)#i
--- xpath
//E[@xml:lang='c' or starts-with(@xml:lang, 'c-')][@id='i']
===
--- selector
E#i:lang(c)
--- xpath
//E[@id='i'][@xml:lang='c' or starts-with(@xml:lang, 'c-')]
===
--- selector
*:lang(c)#i:first-child
--- xpath
//*[@xml:lang='c' or starts-with(@xml:lang, 'c-')][@id='i'][count(preceding-sibling::*) = 0 and parent::*]
===
--- selector
E:lang(c)#i:first-child
--- xpath
//E[@xml:lang='c' or starts-with(@xml:lang, 'c-')][@id='i'][count(preceding-sibling::*) = 0 and parent::*]
===
--- selector
E:lang(c):first-child#i
--- xpath
//E[@xml:lang='c' or starts-with(@xml:lang, 'c-')][count(preceding-sibling::*) = 0 and parent::*][@id='i']
===
--- selector
E#i:lang(c):first-child
--- xpath
//E[@id='i'][@xml:lang='c' or starts-with(@xml:lang, 'c-')][count(preceding-sibling::*) = 0 and parent::*]
===
--- selector
#bar
--- xpath
//*[@id='bar']
===
--- selector
*#bar
--- xpath
//*[@id='bar']
===
--- selector
*[foo]
--- xpath
//*[@foo]
===
--- selector
[foo]
--- xpath
//*[@foo]
===
--- selector
.warning
--- xpath
//*[contains(concat(' ', @class, ' '), ' warning ')]
===
--- selector
*.warning
--- xpath
//*[contains(concat(' ', @class, ' '), ' warning ')]
===
--- selector
:nth-child(1)
--- xpath
//*[count(preceding-sibling::*) = 0 and parent::*]
===
--- selector
*:nth-child(1)
--- xpath
//*[count(preceding-sibling::*) = 0 and parent::*]
===
--- selector
E:nth-child(1)
--- xpath
//E[count(preceding-sibling::*) = 0 and parent::*]
===
--- selector
E:nth-child(2)
--- xpath
//E[count(preceding-sibling::*) = 1 and parent::*]
===
--- selector
:root
--- xpath
/*
===
--- selector
E:root
--- xpath
/E
===
--- selector
E:empty
--- xpath
//E[not(* or text())]
===
--- selector
:empty
--- xpath
//*[not(* or text())]
===
--- selector
p , :root
--- xpath
//p | /*
===
--- selector
p , q
--- xpath
//p | //q
===
--- selector
div *:not(p) em
--- xpath
//div//*[not(self::p)]//em
===
--- selector
div em:only-child
--- xpath
//div//em[count(preceding-sibling::*) = 0 and parent::*][count(following-sibling::*) = 0 and parent::*]
===
--- selector
[x=abc]
--- xpath
//*[@x='abc']
===
--- selector
[x=a-bc]
--- xpath
//*[@x='a-bc']
===
--- selector
[x=abc-]
--- xpath
//*[@x='abc-']
===
--- selector
[x=ab--c]
--- xpath
//*[@x='ab--c']
===
--- selector
option[value!=""]
--- xpath
//option[@value!='']
===
--- selector
option[ value="" ]
--- xpath
//option[@value='']
===
--- selector
tr[class!="wantedClass"]
--- xpath
//tr[@class!='wantedClass']
===
--- selector
form[name='foo']
--- xpath
//form[@name='foo']
HTML-Selector-XPath-0.16/t/99_pod.t 0000644 0001750 0001750 00000000201 12201461127 016070 0 ustar corion corion use Test::More;
eval "use Test::Pod 1.00";
plan skip_all => "Test::Pod 1.00 required for testing POD" if $@;
all_pod_files_ok();
HTML-Selector-XPath-0.16/t/00_compile.t 0000644 0001750 0001750 00000000121 12201461127 016715 0 ustar corion corion use strict;
use Test::More tests => 1;
BEGIN { use_ok 'HTML::Selector::XPath' }
HTML-Selector-XPath-0.16/t/04_subclass.t 0000644 0001750 0001750 00000001345 12201461127 017121 0 ustar corion corion package HTML::Selector::XPath::Extended;
use strict;
use Test::Base;
use base qw(HTML::Selector::XPath);
plan tests => 1 * blocks;
filters { selector => 'chomp', xpath => 'chomp' };
sub parse_pseudo {
my ($self, $pseudo, $expr) = @_;
return q{[@type='foo']} if $pseudo eq 'foo';
return qq{[\@type='bar-$1']} if $pseudo =~ /^bar\(/ && $$expr =~ s/^"([^"]+)"\)//;
return "" if $pseudo eq 'quax';
}
run {
my $block = shift;
my $selector = HTML::Selector::XPath::Extended->new($block->selector);
is $selector->to_xpath, $block->xpath, $block->selector;
}
__END__
===
--- selector
:foo
--- xpath
//*[@type='foo']
===
--- selector
:bar("baz")
--- xpath
//*[@type='bar-baz']
===
--- selector
:quax
--- xpath
//*
HTML-Selector-XPath-0.16/t/05_xpath_prefix.t 0000644 0001750 0001750 00000001543 12201461127 020004 0 ustar corion corion use strict;
use Test::Base;
use HTML::Selector::XPath qw(selector_to_xpath);
plan tests => 1 * blocks;
filters { selector => 'chomp', xpath => 'chomp' };
run {
my $block = shift;
is selector_to_xpath($block->selector, root => '/R', prefix => 'xhtml'), $block->xpath, $block->selector;
}
__END__
===
--- selector
*
--- xpath
/R/*
===
--- selector
E
--- xpath
/R/xhtml:E
===
--- selector
E F
--- xpath
/R/xhtml:E//xhtml:F
===
--- selector
E > F
--- xpath
/R/xhtml:E/xhtml:F
===
--- selector
E + F
--- xpath
/R/xhtml:E/following-sibling::*[1]/self::xhtml:F
===
--- selector
E[foo]
--- xpath
/R/xhtml:E[@foo]
===
--- selector
E[foo="warning"]
--- xpath
/R/xhtml:E[@foo='warning']
===
--- selector
E#myid
--- xpath
/R/xhtml:E[@id='myid']
===
--- selector
foo.bar, bar
--- xpath
/R/xhtml:foo[contains(concat(' ', @class, ' '), ' bar ')] | /R/xhtml:bar
HTML-Selector-XPath-0.16/lib/ 0000755 0001750 0001750 00000000000 12201461131 015105 5 ustar corion corion HTML-Selector-XPath-0.16/lib/HTML/ 0000755 0001750 0001750 00000000000 12201461131 015651 5 ustar corion corion HTML-Selector-XPath-0.16/lib/HTML/Selector/ 0000755 0001750 0001750 00000000000 12201461131 017431 5 ustar corion corion HTML-Selector-XPath-0.16/lib/HTML/Selector/XPath.pm 0000644 0001750 0001750 00000026457 12201461127 021036 0 ustar corion corion package HTML::Selector::XPath;
use strict;
use 5.008_001;
our $VERSION = '0.16';
require Exporter;
our @EXPORT_OK = qw(selector_to_xpath);
*import = \&Exporter::import;
use Carp;
sub selector_to_xpath {
__PACKAGE__->new(shift)->to_xpath(@_);
}
# XXX: Identifiers should also allow any characters U+00A0 and higher, and any
# escaped characters.
my $ident = qr/(?![0-9]|-[-0-9])[-_a-zA-Z0-9]+/;
my $reg = {
# tag name/id/class
element => qr/^([#.]?)([a-z0-9\\*_-]*)((\|)([a-z0-9\\*_-]*))?/i,
# attribute presence
attr1 => qr/^\[ \s* ($ident) \s* \]/x,
# attribute value match
attr2 => qr/^\[ \s* ($ident) \s*
( [~|*^\$!]? = ) \s*
(?: ($ident) | "([^"]*)" | '([^']*)') \s* \] /x,
badattr => qr/^\[/,
attrN => qr/^:not\((.*?)\)/i, # this should be a parentheses matcher instead of a RE!
pseudo => qr/^:([()a-z0-9_+-]+)/i,
# adjacency/direct descendance
combinator => qr/^(\s*[>+~\s](?!,))/i,
# rule separator
comma => qr/^\s*,\s*/i,
};
sub new {
my($class, $exp) = @_;
bless { expression => $exp }, $class;
}
sub selector {
my $self = shift;
$self->{expression} = shift if @_;
$self->{expression};
}
sub convert_attribute_match {
my ($left,$op,$right) = @_;
# negation (e.g. [input!="text"]) isn't implemented in CSS, but include it anyway:
if ($op eq '!=') {
"\@$left!='$right'";
} elsif ($op eq '~=') { # substring attribute match
"contains(concat(' ', \@$left, ' '), ' $right ')";
} elsif ($op eq '*=') { # real substring attribute match
"contains(\@$left, '$right')";
} elsif ($op eq '|=') {
"\@$left='$right' or starts-with(\@$left, '$right-')";
} elsif ($op eq '^=') {
"starts-with(\@$left,'$^N')";
} elsif ($op eq '$=') {
"ends-with(\@$left,'$^N')";
} else { # exact match
"\@$left='$^N'";
}
};
sub _generate_child {
my ($direction,$a,$b) = @_;
if ($a == 0) { # 0n+b
$b--;
"[count($direction-sibling::*) = $b and parent::*]"
} elsif ($a > 0) { # an + b
return "[not((count($direction-sibling::*)+1)<$b) and ((count($direction-sibling::*) + 1) - $b) mod $a = 0 and parent::*]"
} else { # -an + $b
$a = -$a;
return "[not((count($direction-sibling::*)+1)>$b) and (($b - (count($direction-sibling::*) + 1)) mod $a) = 0 and parent::*]"
};
};
sub nth_child {
my ($a,$b) = @_;
if (@_ == 1) {
($a,$b) = (0,$a);
};
_generate_child('preceding', $a, $b);
};
sub nth_last_child {
my ($a,$b) = @_;
if (@_ == 1) {
($a,$b) = (0,$a);
};
_generate_child('following', $a, $b);
};
sub to_xpath {
my $self = shift;
my $rule = $self->{expression} or return;
my %parms = @_;
my $root = $parms{root} || '/';
my @parts = ("$root/");
my $last_rule = '';
my @next_parts;
my $tag;
my $wrote_tag;
my $tag_index;
my $root_index = 0; # points to the current root
# Loop through each "unit" of the rule
while (length $rule && $rule ne $last_rule) {
$last_rule = $rule;
$rule =~ s/^\s*|\s*$//g;
last unless length $rule;
# Prepend explicit first selector if we have an implicit selector
# (that is, if we start with a combinator)
if ($rule =~ /$reg->{combinator}/) {
$rule = "* $rule";
};
# Match elements
if ($rule =~ s/$reg->{element}//) {
my ($id_class,$name,$lang) = ($1,$2,$3);
# to add *[1]/self:: for follow-sibling
if (@next_parts) {
push @parts, @next_parts; #, (pop @parts);
@next_parts = ();
}
if ($id_class eq '') {
$tag = $name || '*';
} else {
$tag = '*';
}
if (defined $parms{prefix} and not $tag =~ /[*:|]/) {
$tag = join ':', $parms{prefix}, $tag;
}
if (! $wrote_tag++) {
push @parts, $tag;
$tag_index = $#parts;
};
# XXX Shouldn't the RE allow both, ID and class?
if ($id_class eq '#') { # ID
push @parts, "[\@id='$name']";
} elsif ($id_class eq '.') { # class
push @parts, "[contains(concat(' ', \@class, ' '), ' $name ')]";
};
};
# Match attribute selectors
if ($rule =~ s/$reg->{attr2}//) {
push @parts, "[", convert_attribute_match( $1, $2, $^N ), "]";
} elsif ($rule =~ s/$reg->{attr1}//) {
# If we have no tag output yet, write the tag:
if (! $wrote_tag++) {
push @parts, '*';
$tag_index = $#parts;
};
push @parts, "[\@$1]";
} elsif ($rule =~ $reg->{badattr}) {
Carp::croak "Invalid attribute-value selector '$rule'";
}
# Match negation
if ($rule =~ s/$reg->{attrN}//) {
my $sub_rule = $1;
if ($sub_rule =~ s/$reg->{attr2}//) {
push @parts, "[not(", convert_attribute_match( $1, $2, $^N ), ")]";
} elsif ($sub_rule =~ s/$reg->{attr1}//) {
push @parts, "[not(\@$1)]";
} elsif ($rule =~ $reg->{badattr}) {
Carp::croak "Invalid attribute-value selector '$rule'";
} else {
my $xpath = selector_to_xpath($sub_rule);
$xpath =~ s!^//!!;
push @parts, "[not(self::$xpath)]";
#} else {
# Carp::croak "Can't translate '$sub_rule' inside :not()";
}
}
# Ignore pseudoclasses/pseudoelements
while ($rule =~ s/$reg->{pseudo}//) {
if ( my @expr = $self->parse_pseudo($1, \$rule) ) {
push @parts, @expr;
} elsif ( $1 eq 'first-child') {
# Translates to :nth-child(1)
push @parts, nth_child(1);
} elsif ( $1 eq 'last-child') {
push @parts, nth_last_child(1);
} elsif ( $1 eq 'only-child') {
push @parts, nth_child(1), nth_last_child(1);
} elsif ($1 =~ /^lang\(([\w\-]+)\)$/) {
push @parts, "[\@xml:lang='$1' or starts-with(\@xml:lang, '$1-')]";
} elsif ($1 =~ /^nth-child\((\d+)\)$/) {
push @parts, nth_child($1);
} elsif ($1 =~ /^nth-child\((\d+)n(?:\+(\d+))?\)$/) {
push @parts, nth_child($1, $2||0);
} elsif ($1 =~ /^nth-last-child\((\d+)\)$/) {
push @parts, nth_last_child($1);
} elsif ($1 =~ /^nth-last-child\((\d+)n(?:\+(\d+))?\)$/) {
push @parts, nth_last_child($1, $2||0);
} elsif ($1 =~ /^first-of-type$/) {
push @parts, "[1]";
} elsif ($1 =~ /^nth-of-type\((\d+)\)$/) {
push @parts, "[$1]";
} elsif ($1 =~ /^contains\($/) {
$rule =~ s/^\s*"([^"]*)"\s*\)//
or die "Malformed string in :contains(): '$rule'";
push @parts, qq{[text()[contains(string(.),"$1")]]};
} elsif ( $1 eq 'root') {
# This will give surprising results if you do E > F:root
$parts[$root_index] = $root;
} elsif ( $1 eq 'empty') {
push @parts, "[not(* or text())]";
} else {
Carp::croak "Can't translate '$1' pseudo-class";
}
}
# Match combinators (whitespace, >, + and ~)
if ($rule =~ s/$reg->{combinator}//) {
my $match = $1;
if ($match =~ />/) {
push @parts, "/";
} elsif ($match =~ /\+/) {
push @parts, "/following-sibling::*[1]/self::";
$tag_index = $#parts;
} elsif ($match =~ /\~/) {
push @parts, "/following-sibling::";
} elsif ($match =~ /^\s*$/) {
push @parts, "//"
} else {
die "Weird combinator '$match'"
}
# new context
undef $tag;
undef $wrote_tag;
}
# Match commas
if ($rule =~ s/$reg->{comma}//) {
push @parts, " | ", "$root/"; # ending one rule and beginning another
$root_index = $#parts;
undef $tag;
undef $wrote_tag;
}
}
return join '', @parts;
}
sub parse_pseudo {
# nop
}
1;
__END__
=head1 NAME
HTML::Selector::XPath - CSS Selector to XPath compiler
=head1 SYNOPSIS
use HTML::Selector::XPath;
my $selector = HTML::Selector::XPath->new("li#main");
$selector->to_xpath; # //li[@id='main']
# functional interface
use HTML::Selector::XPath 'selector_to_xpath';
my $xpath = selector_to_xpath('div.foo');
my $relative = selector_to_xpath('div.foo', root => '/html/body/p' );
# /html/body/p/div[contains(concat(' ', @class, ' '), ' foo ')]
my $relative = selector_to_xpath('div:root', root => '/html/body/p' );
# /html/body/p/div
=head1 DESCRIPTION
HTML::Selector::XPath is a utility function to compile full set of
CSS2 and partial CSS3 selectors to the equivalent XPath expression.
=head1 FUNCTIONS and METHODS
=over 4
=item selector_to_xpath
$xpath = selector_to_xpath($selector, %options);
Shortcut for C<< HTML::Selector->new(shift)->to_xpath(@_) >>. Exported upon request.
=item new
$sel = HTML::Selector::XPath->new($selector, %options);
Creates a new object.
=item to_xpath
$xpath = $sel->to_xpath;
$xpath = $sel->to_xpath(root => "."); # ./foo instead of //foo
Returns the translated XPath expression. You can optionally pass
C
parameter, to specify which root to start the expression. It
defaults to C.
The optional C option allows you to specify a namespace
prefix for the generated XPath expression.
=back
=head1 SUBCLASSING NOTES
=over 4
=item parse_pseudo
This method is called during xpath construction when we encounter a pseudo
selector (something that begins with comma). It is passed the selector and
a reference to the string we are parsing. It should return one or more
xpath sub-expressions to add to the parts if the selector is handled,
otherwise return an empty list.
=back
=head1 CAVEATS
=head2 CSS SELECTOR VALIDATION
This module doesn't validate whether the original CSS Selector
expression is valid. For example,
div.123foo
is an invalid CSS selector (class names should not begin with
numbers), but this module ignores that and tries to generate
an equivalent XPath expression anyway.
=head1 COPYRIGHT
Tatsuhiko Miyagawa 2006-2011
Max Maischein 2011-
=head1 AUTHOR
Tatsuhiko Miyagawa Emiyagawa@bulknews.netE
Most of the logic is based on Joe Hewitt's getElementsBySelector.js on
L and Andrew Dupont's
patch to Prototype.js on L,
but slightly modified using Aristotle Pegaltzis' CSS to XPath
translation table per L
Also see
L
and
L
=head1 LICENSE
This library is free software; you can redistribute it and/or modify
it under the same terms as Perl itself.
=head1 SEE ALSO
L
L
=cut