mirror of
https://github.com/ruby/ruby.git
synced 2022-11-09 12:17:21 -05:00
Cross-ported the REXML changes (3.0.8) from the development branch to the
stable branch. git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_1_8@6339 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
This commit is contained in:
parent
db0fac0266
commit
abe1214b3d
12 changed files with 946 additions and 891 deletions
|
@ -3,309 +3,308 @@ require 'rexml/xmltokens'
|
|||
require 'rexml/parsers/xpathparser'
|
||||
|
||||
module REXML
|
||||
# You don't want to use this class. Really. Use XPath, which is a wrapper
|
||||
# for this class. Believe me. You don't want to poke around in here.
|
||||
# There is strange, dark magic at work in this code. Beware. Go back! Go
|
||||
# back while you still can!
|
||||
class XPathParser
|
||||
include XMLTokens
|
||||
LITERAL = /^'([^']*)'|^"([^"]*)"/u
|
||||
# You don't want to use this class. Really. Use XPath, which is a wrapper
|
||||
# for this class. Believe me. You don't want to poke around in here.
|
||||
# There is strange, dark magic at work in this code. Beware. Go back! Go
|
||||
# back while you still can!
|
||||
class XPathParser
|
||||
include XMLTokens
|
||||
LITERAL = /^'([^']*)'|^"([^"]*)"/u
|
||||
|
||||
def initialize( )
|
||||
@parser = REXML::Parsers::XPathParser.new
|
||||
@namespaces = {}
|
||||
@variables = {}
|
||||
end
|
||||
def initialize( )
|
||||
@parser = REXML::Parsers::XPathParser.new
|
||||
@namespaces = {}
|
||||
@variables = {}
|
||||
end
|
||||
|
||||
def namespaces=( namespaces={} )
|
||||
Functions::namespace_context = namespaces
|
||||
@namespaces = namespaces
|
||||
end
|
||||
def namespaces=( namespaces={} )
|
||||
Functions::namespace_context = namespaces
|
||||
@namespaces = namespaces
|
||||
end
|
||||
|
||||
def variables=( vars={} )
|
||||
Functions::variables = vars
|
||||
@variables = vars
|
||||
end
|
||||
def variables=( vars={} )
|
||||
Functions::variables = vars
|
||||
@variables = vars
|
||||
end
|
||||
|
||||
def parse path, nodeset
|
||||
path_stack = @parser.parse( path )
|
||||
def parse path, nodeset
|
||||
path_stack = @parser.parse( path )
|
||||
#puts "PARSE: #{path} => #{path_stack.inspect}"
|
||||
#puts "PARSE: nodeset = #{nodeset.collect{|x|x.to_s}.inspect}"
|
||||
match( path_stack, nodeset )
|
||||
end
|
||||
match( path_stack, nodeset )
|
||||
end
|
||||
|
||||
def predicate path, nodeset
|
||||
path_stack = @parser.predicate( path )
|
||||
return Predicate( path_stack, nodeset )
|
||||
end
|
||||
def predicate path, nodeset
|
||||
path_stack = @parser.predicate( path )
|
||||
return Predicate( path_stack, nodeset )
|
||||
end
|
||||
|
||||
def []=( variable_name, value )
|
||||
@variables[ variable_name ] = value
|
||||
end
|
||||
def []=( variable_name, value )
|
||||
@variables[ variable_name ] = value
|
||||
end
|
||||
|
||||
private
|
||||
private
|
||||
|
||||
def match( path_stack, nodeset )
|
||||
while ( path_stack.size > 0 and nodeset.size > 0 )
|
||||
#puts "PARSE: #{path_stack.inspect} '#{nodeset.collect{|n|n.class}.inspect}'"
|
||||
nodeset = internal_parse( path_stack, nodeset )
|
||||
#puts "NODESET: #{nodeset.size}"
|
||||
#puts "PATH_STACK: #{path_stack.inspect}"
|
||||
end
|
||||
nodeset
|
||||
end
|
||||
def match( path_stack, nodeset )
|
||||
while ( path_stack.size > 0 and nodeset.size > 0 )
|
||||
#puts "PARSE: #{path_stack.inspect} '#{nodeset.collect{|n|n.class}.inspect}'"
|
||||
nodeset = internal_parse( path_stack, nodeset )
|
||||
#puts "NODESET: #{nodeset}"
|
||||
#puts "PATH_STACK: #{path_stack.inspect}"
|
||||
end
|
||||
nodeset
|
||||
end
|
||||
|
||||
def internal_parse path_stack, nodeset
|
||||
def internal_parse path_stack, nodeset
|
||||
#puts "INTERNAL_PARSE RETURNING WITH NO RESULTS" if nodeset.size == 0 or path_stack.size == 0
|
||||
return nodeset if nodeset.size == 0 or path_stack.size == 0
|
||||
#puts "INTERNAL_PARSE: #{path_stack.inspect}, #{nodeset.collect{|n| n.class}.inspect}"
|
||||
case path_stack.shift
|
||||
when :document
|
||||
return [ nodeset[0].root.parent ]
|
||||
return nodeset if nodeset.size == 0 or path_stack.size == 0
|
||||
#puts "INTERNAL_PARSE: #{path_stack.inspect}, #{nodeset.collect{|n| n.class}.inspect}"
|
||||
case path_stack.shift
|
||||
when :document
|
||||
return [ nodeset[0].root.parent ]
|
||||
|
||||
when :qname
|
||||
prefix = path_stack.shift
|
||||
name = path_stack.shift
|
||||
#puts "QNAME #{prefix}#{prefix.size>0?':':''}#{name}"
|
||||
n = nodeset.clone
|
||||
ns = @namespaces[prefix]
|
||||
ns = ns ? ns : ''
|
||||
n.delete_if do |node|
|
||||
# FIXME: This DOUBLES the time XPath searches take
|
||||
ns = node.namespace( prefix ) if node.node_type == :element and ns == ''
|
||||
#puts "NODE: '#{node.to_s}'; node.has_name?( #{name.inspect}, #{ns.inspect} ): #{ node.has_name?( name, ns )}; node.namespace() = #{node.namespace().inspect}; node.prefix = #{node.prefix().inspect}" if node.node_type == :element
|
||||
!(node.node_type == :element and node.name == name and node.namespace == ns )
|
||||
end
|
||||
return n
|
||||
when :qname
|
||||
prefix = path_stack.shift
|
||||
name = path_stack.shift
|
||||
#puts "QNAME #{prefix}#{prefix.size>0?':':''}#{name}"
|
||||
n = nodeset.clone
|
||||
ns = @namespaces[prefix]
|
||||
ns = ns ? ns : ''
|
||||
n.delete_if do |node|
|
||||
# FIXME: This DOUBLES the time XPath searches take
|
||||
ns = node.namespace( prefix ) if node.node_type == :element and ns == ''
|
||||
#puts "NODE: '#{node.to_s}'; node.has_name?( #{name.inspect}, #{ns.inspect} ): #{ node.has_name?( name, ns )}; node.namespace() = #{node.namespace().inspect}; node.prefix = #{node.prefix().inspect}" if node.node_type == :element
|
||||
!(node.node_type == :element and node.name == name and node.namespace == ns )
|
||||
end
|
||||
return n
|
||||
|
||||
when :any
|
||||
n = nodeset.clone
|
||||
n.delete_if { |node| node.node_type != :element }
|
||||
return n
|
||||
when :any
|
||||
n = nodeset.clone
|
||||
n.delete_if { |node| node.node_type != :element }
|
||||
return n
|
||||
|
||||
when :self
|
||||
# THIS SPACE LEFT INTENTIONALLY BLANK
|
||||
when :self
|
||||
# THIS SPACE LEFT INTENTIONALLY BLANK
|
||||
|
||||
when :processing_instruction
|
||||
target = path_stack.shift
|
||||
n = nodeset.clone
|
||||
n.delete_if do |node|
|
||||
(node.node_type != :processing_instruction) or
|
||||
( !target.nil? and ( node.target != target ) )
|
||||
end
|
||||
return n
|
||||
when :processing_instruction
|
||||
target = path_stack.shift
|
||||
n = nodeset.clone
|
||||
n.delete_if do |node|
|
||||
(node.node_type != :processing_instruction) or
|
||||
( !target.nil? and ( node.target != target ) )
|
||||
end
|
||||
return n
|
||||
|
||||
when :text
|
||||
#puts ":TEXT"
|
||||
n = nodeset.clone
|
||||
n.delete_if do |node|
|
||||
#puts "#{node} :: #{node.node_type}"
|
||||
node.node_type != :text
|
||||
end
|
||||
return n
|
||||
when :text
|
||||
#puts ":TEXT"
|
||||
n = nodeset.clone
|
||||
n.delete_if do |node|
|
||||
#puts "#{node} :: #{node.node_type}"
|
||||
node.node_type != :text
|
||||
end
|
||||
return n
|
||||
|
||||
when :comment
|
||||
n = nodeset.clone
|
||||
n.delete_if do |node|
|
||||
node.node_type != :comment
|
||||
end
|
||||
return n
|
||||
when :comment
|
||||
n = nodeset.clone
|
||||
n.delete_if do |node|
|
||||
node.node_type != :comment
|
||||
end
|
||||
return n
|
||||
|
||||
when :node
|
||||
return nodeset
|
||||
|
||||
# FIXME: I suspect the following XPath will fail:
|
||||
# /a/*/*[1]
|
||||
when :child
|
||||
#puts "CHILD"
|
||||
new_nodeset = []
|
||||
nt = nil
|
||||
for node in nodeset
|
||||
nt = node.node_type
|
||||
new_nodeset += node.children if nt == :element or nt == :document
|
||||
end
|
||||
#path_stack[0,(path_stack.size-ps_clone.size)] = []
|
||||
return new_nodeset
|
||||
when :node
|
||||
return nodeset
|
||||
|
||||
# FIXME: I suspect the following XPath will fail:
|
||||
# /a/*/*[1]
|
||||
when :child
|
||||
#puts "CHILD"
|
||||
new_nodeset = []
|
||||
nt = nil
|
||||
for node in nodeset
|
||||
nt = node.node_type
|
||||
new_nodeset += node.children if nt == :element or nt == :document
|
||||
end
|
||||
#path_stack[0,(path_stack.size-ps_clone.size)] = []
|
||||
return new_nodeset
|
||||
|
||||
when :literal
|
||||
literal = path_stack.shift
|
||||
if literal =~ /^\d+(\.\d+)?$/
|
||||
return ($1 ? literal.to_f : literal.to_i)
|
||||
end
|
||||
#puts "RETURNING '#{literal}'"
|
||||
return literal
|
||||
|
||||
when :attribute
|
||||
#puts ":ATTRIBUTE"
|
||||
new_nodeset = []
|
||||
case path_stack.shift
|
||||
when :qname
|
||||
prefix = path_stack.shift
|
||||
name = path_stack.shift
|
||||
for element in nodeset
|
||||
if element.node_type == :element
|
||||
#puts element.name
|
||||
#puts "looking for attribute #{name} in '#{@namespaces[prefix]}'"
|
||||
attr = element.attribute( name, @namespaces[prefix] )
|
||||
#puts ":ATTRIBUTE: attr => #{attr}"
|
||||
new_nodeset << attr if attr
|
||||
end
|
||||
end
|
||||
when :any
|
||||
for element in nodeset
|
||||
if element.node_type == :element
|
||||
attr = element.attributes
|
||||
end
|
||||
end
|
||||
end
|
||||
#puts "RETURNING #{new_nodeset.collect{|n|n.to_s}.inspect}"
|
||||
return new_nodeset
|
||||
when :literal
|
||||
literal = path_stack.shift
|
||||
if literal =~ /^\d+(\.\d+)?$/
|
||||
return ($1 ? literal.to_f : literal.to_i)
|
||||
end
|
||||
#puts "RETURNING '#{literal}'"
|
||||
return literal
|
||||
|
||||
when :attribute
|
||||
new_nodeset = []
|
||||
case path_stack.shift
|
||||
when :qname
|
||||
prefix = path_stack.shift
|
||||
name = path_stack.shift
|
||||
for element in nodeset
|
||||
if element.node_type == :element
|
||||
#puts element.name
|
||||
#puts "looking for attribute #{name} in '#{@namespaces[prefix]}'"
|
||||
attr = element.attribute( name, @namespaces[prefix] )
|
||||
#puts ":ATTRIBUTE: attr => #{attr}"
|
||||
new_nodeset << attr if attr
|
||||
end
|
||||
end
|
||||
when :any
|
||||
for element in nodeset
|
||||
if element.node_type == :element
|
||||
attr = element.attributes
|
||||
end
|
||||
end
|
||||
end
|
||||
#puts "RETURNING #{new_nodeset.collect{|n|n.to_s}.inspect}"
|
||||
return new_nodeset
|
||||
|
||||
when :parent
|
||||
return internal_parse( path_stack, nodeset.collect{|n| n.parent}.compact )
|
||||
when :parent
|
||||
return internal_parse( path_stack, nodeset.collect{|n| n.parent}.compact )
|
||||
|
||||
when :ancestor
|
||||
#puts "ANCESTOR"
|
||||
new_nodeset = []
|
||||
for node in nodeset
|
||||
while node.parent
|
||||
node = node.parent
|
||||
new_nodeset << node unless new_nodeset.include? node
|
||||
end
|
||||
end
|
||||
#nodeset = new_nodeset.uniq
|
||||
return new_nodeset
|
||||
when :ancestor
|
||||
#puts "ANCESTOR"
|
||||
new_nodeset = []
|
||||
for node in nodeset
|
||||
while node.parent
|
||||
node = node.parent
|
||||
new_nodeset << node unless new_nodeset.include? node
|
||||
end
|
||||
end
|
||||
#nodeset = new_nodeset.uniq
|
||||
return new_nodeset
|
||||
|
||||
when :ancestor_or_self
|
||||
new_nodeset = []
|
||||
for node in nodeset
|
||||
if node.node_type == :element
|
||||
new_nodeset << node
|
||||
while ( node.parent )
|
||||
node = node.parent
|
||||
new_nodeset << node unless new_nodeset.includes? node
|
||||
end
|
||||
end
|
||||
end
|
||||
#nodeset = new_nodeset.uniq
|
||||
return new_nodeset
|
||||
when :ancestor_or_self
|
||||
new_nodeset = []
|
||||
for node in nodeset
|
||||
if node.node_type == :element
|
||||
new_nodeset << node
|
||||
while ( node.parent )
|
||||
node = node.parent
|
||||
new_nodeset << node unless new_nodeset.includes? node
|
||||
end
|
||||
end
|
||||
end
|
||||
#nodeset = new_nodeset.uniq
|
||||
return new_nodeset
|
||||
|
||||
when :predicate
|
||||
#puts "@"*80
|
||||
#puts "NODESET = #{nodeset.collect{|n|n.to_s}.inspect}"
|
||||
predicate = path_stack.shift
|
||||
new_nodeset = []
|
||||
Functions::size = nodeset.size
|
||||
nodeset.size.times do |index|
|
||||
node = nodeset[index]
|
||||
Functions::node = node
|
||||
Functions::index = index+1
|
||||
#puts "Node #{node} and index=#{index+1}"
|
||||
result = Predicate( predicate, node )
|
||||
#puts "Predicate returned #{result} (#{result.class}) for #{node.class}"
|
||||
if result.kind_of? Numeric
|
||||
#puts "#{result} == #{index} => #{result == index}"
|
||||
new_nodeset << node if result == (index+1)
|
||||
elsif result.instance_of? Array
|
||||
new_nodeset << node if result.size > 0
|
||||
else
|
||||
new_nodeset << node if result
|
||||
end
|
||||
end
|
||||
#puts "Nodeset after predicate #{predicate.inspect} has #{new_nodeset.size} nodes"
|
||||
#puts "NODESET: #{new_nodeset.collect{|n|n.to_s}.inspect}"
|
||||
return new_nodeset
|
||||
when :predicate
|
||||
#puts "@"*80
|
||||
#puts "NODESET = #{nodeset.collect{|n|n.to_s}.inspect}"
|
||||
predicate = path_stack.shift
|
||||
new_nodeset = []
|
||||
Functions::size = nodeset.size
|
||||
nodeset.size.times do |index|
|
||||
node = nodeset[index]
|
||||
Functions::node = node
|
||||
Functions::index = index+1
|
||||
#puts "Node #{node} and index=#{index+1}"
|
||||
result = Predicate( predicate, node )
|
||||
#puts "Predicate returned #{result} (#{result.class}) for #{node.class}"
|
||||
if result.kind_of? Numeric
|
||||
#puts "#{result} == #{index} => #{result == index}"
|
||||
new_nodeset << node if result == (index+1)
|
||||
elsif result.instance_of? Array
|
||||
new_nodeset << node if result.size > 0
|
||||
else
|
||||
new_nodeset << node if result
|
||||
end
|
||||
end
|
||||
#puts "Nodeset after predicate #{predicate.inspect} has #{new_nodeset.size} nodes"
|
||||
#puts "NODESET: #{new_nodeset.collect{|n|n.to_s}.inspect}"
|
||||
return new_nodeset
|
||||
|
||||
when :descendant_or_self
|
||||
rv = descendant_or_self( path_stack, nodeset )
|
||||
path_stack.clear
|
||||
return rv
|
||||
when :descendant_or_self
|
||||
rv = descendant_or_self( path_stack, nodeset )
|
||||
path_stack.clear
|
||||
return rv
|
||||
|
||||
when :descendant
|
||||
#puts ":DESCENDANT"
|
||||
results = []
|
||||
nt = nil
|
||||
for node in nodeset
|
||||
nt = node.node_type
|
||||
results += internal_parse( path_stack.clone.unshift( :descendant_or_self ),
|
||||
node.children ) if nt == :element or nt == :document
|
||||
end
|
||||
return results
|
||||
when :descendant
|
||||
#puts ":DESCENDANT"
|
||||
results = []
|
||||
nt = nil
|
||||
for node in nodeset
|
||||
nt = node.node_type
|
||||
results += internal_parse( path_stack.clone.unshift( :descendant_or_self ),
|
||||
node.children ) if nt == :element or nt == :document
|
||||
end
|
||||
return results
|
||||
|
||||
when :following_sibling
|
||||
results = []
|
||||
for node in nodeset
|
||||
all_siblings = node.parent.children
|
||||
current_index = all_siblings.index( node )
|
||||
following_siblings = all_siblings[ current_index+1 .. -1 ]
|
||||
results += internal_parse( path_stack.clone, following_siblings )
|
||||
end
|
||||
return results
|
||||
when :following_sibling
|
||||
results = []
|
||||
for node in nodeset
|
||||
all_siblings = node.parent.children
|
||||
current_index = all_siblings.index( node )
|
||||
following_siblings = all_siblings[ current_index+1 .. -1 ]
|
||||
results += internal_parse( path_stack.clone, following_siblings )
|
||||
end
|
||||
return results
|
||||
|
||||
when :preceding_sibling
|
||||
results = []
|
||||
for node in nodeset
|
||||
all_siblings = node.parent.children
|
||||
current_index = all_siblings.index( node )
|
||||
preceding_siblings = all_siblings[ 0 .. current_index-1 ]
|
||||
results += internal_parse( path_stack.clone, preceding_siblings )
|
||||
end
|
||||
return results
|
||||
when :preceding_sibling
|
||||
results = []
|
||||
for node in nodeset
|
||||
all_siblings = node.parent.children
|
||||
current_index = all_siblings.index( node )
|
||||
preceding_siblings = all_siblings[ 0 .. current_index-1 ]
|
||||
results += internal_parse( path_stack.clone, preceding_siblings )
|
||||
end
|
||||
return results
|
||||
|
||||
when :preceding
|
||||
new_nodeset = []
|
||||
for node in nodeset
|
||||
new_nodeset += preceding( node )
|
||||
end
|
||||
return new_nodeset
|
||||
when :preceding
|
||||
new_nodeset = []
|
||||
for node in nodeset
|
||||
new_nodeset += preceding( node )
|
||||
end
|
||||
return new_nodeset
|
||||
|
||||
when :following
|
||||
new_nodeset = []
|
||||
for node in nodeset
|
||||
new_nodeset += following( node )
|
||||
end
|
||||
return new_nodeset
|
||||
when :following
|
||||
new_nodeset = []
|
||||
for node in nodeset
|
||||
new_nodeset += following( node )
|
||||
end
|
||||
return new_nodeset
|
||||
|
||||
when :namespace
|
||||
new_set = []
|
||||
for node in nodeset
|
||||
new_nodeset << node.namespace if node.node_type == :element or node.node_type == :attribute
|
||||
end
|
||||
return new_nodeset
|
||||
when :namespace
|
||||
new_set = []
|
||||
for node in nodeset
|
||||
new_nodeset << node.namespace if node.node_type == :element or node.node_type == :attribute
|
||||
end
|
||||
return new_nodeset
|
||||
|
||||
when :variable
|
||||
var_name = path_stack.shift
|
||||
return @variables[ var_name ]
|
||||
when :variable
|
||||
var_name = path_stack.shift
|
||||
return @variables[ var_name ]
|
||||
|
||||
end
|
||||
nodeset
|
||||
end
|
||||
end
|
||||
nodeset
|
||||
end
|
||||
|
||||
##########################################################
|
||||
##########################################################
|
||||
# FIXME
|
||||
# The next two methods are BAD MOJO!
|
||||
# This is my achilles heel. If anybody thinks of a better
|
||||
# way of doing this, be my guest. This really sucks, but
|
||||
# it took me three days to get it to work at all.
|
||||
# ########################################################
|
||||
|
||||
def descendant_or_self( path_stack, nodeset )
|
||||
rs = []
|
||||
d_o_s( path_stack, nodeset, rs )
|
||||
# The next two methods are BAD MOJO!
|
||||
# This is my achilles heel. If anybody thinks of a better
|
||||
# way of doing this, be my guest. This really sucks, but
|
||||
# it took me three days to get it to work at all.
|
||||
# ########################################################
|
||||
|
||||
def descendant_or_self( path_stack, nodeset )
|
||||
rs = []
|
||||
d_o_s( path_stack, nodeset, rs )
|
||||
#puts "RS = #{rs.collect{|n|n.to_s}.inspect}"
|
||||
document_order(rs.flatten.compact)
|
||||
end
|
||||
end
|
||||
|
||||
def d_o_s( p, ns, r )
|
||||
nt = nil
|
||||
ns.each_index do |i|
|
||||
n = ns[i]
|
||||
x = match( p.clone, [ n ] )
|
||||
nt = n.node_type
|
||||
d_o_s( p, n.children, x ) if nt == :element or nt == :document and n.children.size > 0
|
||||
def d_o_s( p, ns, r )
|
||||
nt = nil
|
||||
ns.each_index do |i|
|
||||
n = ns[i]
|
||||
x = match( p.clone, [ n ] )
|
||||
nt = n.node_type
|
||||
d_o_s( p, n.children, x ) if nt == :element or nt == :document and n.children.size > 0
|
||||
r.concat(x) if x.size > 0
|
||||
end
|
||||
end
|
||||
end
|
||||
end
|
||||
|
||||
|
||||
# Reorders an array of nodes so that they are in document order
|
||||
|
@ -327,221 +326,231 @@ module REXML
|
|||
|
||||
def recurse( nodeset, &block )
|
||||
for node in nodeset
|
||||
yield node
|
||||
yield node
|
||||
recurse( node, &block ) if node.node_type == :element
|
||||
end
|
||||
end
|
||||
|
||||
|
||||
# Given a predicate, a node, and a context, evaluates to true or false.
|
||||
def Predicate( predicate, node )
|
||||
predicate = predicate.clone
|
||||
#puts "#"*20
|
||||
#puts "Predicate( #{predicate.inspect}, #{node.class} )"
|
||||
results = []
|
||||
case (predicate[0])
|
||||
when :and, :or, :eq, :neq, :lt, :lteq, :gt, :gteq
|
||||
eq = predicate.shift
|
||||
left = Predicate( predicate.shift, node )
|
||||
right = Predicate( predicate.shift, node )
|
||||
return equality_relational_compare( left, eq, right )
|
||||
# Given a predicate, a node, and a context, evaluates to true or false.
|
||||
def Predicate( predicate, node )
|
||||
predicate = predicate.clone
|
||||
#puts "#"*20
|
||||
#puts "Predicate( #{predicate.inspect}, #{node.class} )"
|
||||
results = []
|
||||
case (predicate[0])
|
||||
when :and, :or, :eq, :neq, :lt, :lteq, :gt, :gteq
|
||||
eq = predicate.shift
|
||||
left = Predicate( predicate.shift, node )
|
||||
right = Predicate( predicate.shift, node )
|
||||
#puts "LEFT = #{left.inspect}"
|
||||
#puts "RIGHT = #{right.inspect}"
|
||||
return equality_relational_compare( left, eq, right )
|
||||
|
||||
when :div, :mod, :mult, :plus, :minus
|
||||
op = predicate.shift
|
||||
left = Predicate( predicate.shift, node )
|
||||
right = Predicate( predicate.shift, node )
|
||||
left = Functions::number( left )
|
||||
right = Functions::number( right )
|
||||
case op
|
||||
when :div
|
||||
return left.to_f / right.to_f
|
||||
when :mod
|
||||
return left % right
|
||||
when :mult
|
||||
return left * right
|
||||
when :plus
|
||||
return left + right
|
||||
when :minus
|
||||
return left - right
|
||||
end
|
||||
when :div, :mod, :mult, :plus, :minus
|
||||
op = predicate.shift
|
||||
left = Predicate( predicate.shift, node )
|
||||
right = Predicate( predicate.shift, node )
|
||||
#puts "LEFT = #{left.inspect}"
|
||||
#puts "RIGHT = #{right.inspect}"
|
||||
left = Functions::number( left )
|
||||
right = Functions::number( right )
|
||||
#puts "LEFT = #{left.inspect}"
|
||||
#puts "RIGHT = #{right.inspect}"
|
||||
case op
|
||||
when :div
|
||||
return left.to_f / right.to_f
|
||||
when :mod
|
||||
return left % right
|
||||
when :mult
|
||||
return left * right
|
||||
when :plus
|
||||
return left + right
|
||||
when :minus
|
||||
return left - right
|
||||
end
|
||||
|
||||
when :union
|
||||
predicate.shift
|
||||
left = Predicate( predicate.shift, node )
|
||||
right = Predicate( predicate.shift, node )
|
||||
left = Predicate( predicate.shift, node )
|
||||
right = Predicate( predicate.shift, node )
|
||||
return (left | right)
|
||||
|
||||
when :neg
|
||||
predicate.shift
|
||||
operand = Functions::number(Predicate( predicate, node ))
|
||||
return -operand
|
||||
when :neg
|
||||
predicate.shift
|
||||
operand = Functions::number(Predicate( predicate, node ))
|
||||
return -operand
|
||||
|
||||
when :not
|
||||
predicate.shift
|
||||
return !Predicate( predicate.shift, node )
|
||||
when :not
|
||||
predicate.shift
|
||||
return !Predicate( predicate.shift, node )
|
||||
|
||||
when :function
|
||||
predicate.shift
|
||||
func_name = predicate.shift.tr('-', '_')
|
||||
arguments = predicate.shift
|
||||
#puts "\nFUNCTION: #{func_name}"
|
||||
#puts "ARGUMENTS: #{arguments.inspect} #{node.to_s}"
|
||||
args = arguments.collect { |arg| Predicate( arg, node ) }
|
||||
#puts "FUNCTION: #{func_name}( #{args.collect{|n|n.to_s}.inspect} )"
|
||||
result = Functions.send( func_name, *args )
|
||||
#puts "RESULTS: #{result.inspect}"
|
||||
return result
|
||||
when :function
|
||||
predicate.shift
|
||||
func_name = predicate.shift.tr('-', '_')
|
||||
arguments = predicate.shift
|
||||
#puts "\nFUNCTION: #{func_name}"
|
||||
#puts "ARGUMENTS: #{arguments.inspect} #{node.to_s}"
|
||||
args = arguments.collect { |arg| Predicate( arg, node ) }
|
||||
#puts "FUNCTION: #{func_name}( #{args.collect{|n|n.to_s}.inspect} )"
|
||||
result = Functions.send( func_name, *args )
|
||||
#puts "RESULTS: #{result.inspect}"
|
||||
return result
|
||||
|
||||
else
|
||||
return match( predicate, [ node ] )
|
||||
else
|
||||
return match( predicate, [ node ] )
|
||||
|
||||
end
|
||||
end
|
||||
end
|
||||
end
|
||||
|
||||
# Builds a nodeset of all of the following nodes of the supplied node,
|
||||
# in document order
|
||||
def following( node )
|
||||
all_siblings = node.parent.children
|
||||
current_index = all_siblings.index( node )
|
||||
following_siblings = all_siblings[ current_index+1 .. -1 ]
|
||||
following = []
|
||||
recurse( following_siblings ) { |node| following << node }
|
||||
following.shift
|
||||
#puts "following is returning #{puta following}"
|
||||
following
|
||||
end
|
||||
# Builds a nodeset of all of the following nodes of the supplied node,
|
||||
# in document order
|
||||
def following( node )
|
||||
all_siblings = node.parent.children
|
||||
current_index = all_siblings.index( node )
|
||||
following_siblings = all_siblings[ current_index+1 .. -1 ]
|
||||
following = []
|
||||
recurse( following_siblings ) { |node| following << node }
|
||||
following.shift
|
||||
#puts "following is returning #{puta following}"
|
||||
following
|
||||
end
|
||||
|
||||
# Builds a nodeset of all of the preceding nodes of the supplied node,
|
||||
# in reverse document order
|
||||
def preceding( node )
|
||||
all_siblings = node.parent.children
|
||||
current_index = all_siblings.index( node )
|
||||
preceding_siblings = all_siblings[ 0 .. current_index-1 ]
|
||||
# Builds a nodeset of all of the preceding nodes of the supplied node,
|
||||
# in reverse document order
|
||||
def preceding( node )
|
||||
all_siblings = node.parent.children
|
||||
current_index = all_siblings.index( node )
|
||||
preceding_siblings = all_siblings[ 0 .. current_index-1 ]
|
||||
|
||||
preceding_siblings.reverse!
|
||||
preceding = []
|
||||
recurse( preceding_siblings ) { |node| preceding << node }
|
||||
preceding.reverse
|
||||
end
|
||||
preceding_siblings.reverse!
|
||||
preceding = []
|
||||
recurse( preceding_siblings ) { |node| preceding << node }
|
||||
preceding.reverse
|
||||
end
|
||||
|
||||
def equality_relational_compare( set1, op, set2 )
|
||||
def equality_relational_compare( set1, op, set2 )
|
||||
#puts "EQ_REL_COMP: #{set1.to_s}, #{op}, #{set2.to_s}"
|
||||
#puts "#{set1.class.name} #{op} #{set2.class.name}"
|
||||
if set1.kind_of? Array and set2.kind_of? Array
|
||||
if set1.kind_of? Array and set2.kind_of? Array
|
||||
#puts "#{set1.size} & #{set2.size}"
|
||||
if set1.size == 1 and set2.size == 1
|
||||
set1 = set1[0]
|
||||
set2 = set2[0]
|
||||
if set1.size == 1 and set2.size == 1
|
||||
set1 = set1[0]
|
||||
set2 = set2[0]
|
||||
elsif set1.size == 0 or set2.size == 0
|
||||
nd = set1.size==0 ? set2 : set1
|
||||
nd.each { |il| return true if compare( il, op, nil ) }
|
||||
else
|
||||
set1.each do |i1|
|
||||
i1 = i1.to_s
|
||||
set2.each do |i2|
|
||||
i2 = i2.to_s
|
||||
return true if compare( i1, op, i2 )
|
||||
end
|
||||
end
|
||||
return false
|
||||
end
|
||||
end
|
||||
else
|
||||
set1.each do |i1|
|
||||
i1 = i1.to_s
|
||||
set2.each do |i2|
|
||||
i2 = i2.to_s
|
||||
return true if compare( i1, op, i2 )
|
||||
end
|
||||
end
|
||||
return false
|
||||
end
|
||||
end
|
||||
#puts "COMPARING VALUES"
|
||||
# If one is nodeset and other is number, compare number to each item
|
||||
# in nodeset s.t. number op number(string(item))
|
||||
# If one is nodeset and other is string, compare string to each item
|
||||
# in nodeset s.t. string op string(item)
|
||||
# If one is nodeset and other is boolean, compare boolean to each item
|
||||
# in nodeset s.t. boolean op boolean(item)
|
||||
if set1.kind_of? Array or set2.kind_of? Array
|
||||
# If one is nodeset and other is number, compare number to each item
|
||||
# in nodeset s.t. number op number(string(item))
|
||||
# If one is nodeset and other is string, compare string to each item
|
||||
# in nodeset s.t. string op string(item)
|
||||
# If one is nodeset and other is boolean, compare boolean to each item
|
||||
# in nodeset s.t. boolean op boolean(item)
|
||||
if set1.kind_of? Array or set2.kind_of? Array
|
||||
#puts "ISA ARRAY"
|
||||
if set1.kind_of? Array
|
||||
a = set1
|
||||
b = set2.to_s
|
||||
else
|
||||
a = set2
|
||||
b = set1.to_s
|
||||
end
|
||||
if set1.kind_of? Array
|
||||
a = set1
|
||||
b = set2.to_s
|
||||
else
|
||||
a = set2
|
||||
b = set1.to_s
|
||||
end
|
||||
|
||||
case b
|
||||
when 'true', 'false'
|
||||
b = Functions::boolean( b )
|
||||
for v in a
|
||||
v = Functions::boolean(v)
|
||||
return true if compare( v, op, b )
|
||||
end
|
||||
when /^\d+(\.\d+)?$/
|
||||
b = Functions::number( b )
|
||||
for v in a
|
||||
v = Functions::number(v)
|
||||
return true if compare( v, op, b )
|
||||
end
|
||||
else
|
||||
b = Functions::string( b )
|
||||
for v in a
|
||||
v = Functions::string(v)
|
||||
return true if compare( v, op, b )
|
||||
end
|
||||
end
|
||||
else
|
||||
# If neither is nodeset,
|
||||
# If op is = or !=
|
||||
# If either boolean, convert to boolean
|
||||
# If either number, convert to number
|
||||
# Else, convert to string
|
||||
# Else
|
||||
# Convert both to numbers and compare
|
||||
s1 = set1.to_s
|
||||
s2 = set2.to_s
|
||||
#puts "EQ_REL_COMP: #{set1}=>#{s1}, #{set2}=>#{s2}"
|
||||
if s1 == 'true' or s1 == 'false' or s2 == 'true' or s2 == 'false'
|
||||
#puts "Functions::boolean(#{set1})=>#{Functions::boolean(set1)}"
|
||||
#puts "Functions::boolean(#{set2})=>#{Functions::boolean(set2)}"
|
||||
set1 = Functions::boolean( set1 )
|
||||
set2 = Functions::boolean( set2 )
|
||||
else
|
||||
if op == :eq or op == :neq
|
||||
if s1 =~ /^\d+(\.\d+)?$/ or s2 =~ /^\d+(\.\d+)?$/
|
||||
set1 = Functions::number( s1 )
|
||||
set2 = Functions::number( s2 )
|
||||
else
|
||||
set1 = Functions::string( set1 )
|
||||
set2 = Functions::string( set2 )
|
||||
end
|
||||
else
|
||||
set1 = Functions::number( set1 )
|
||||
set2 = Functions::number( set2 )
|
||||
end
|
||||
end
|
||||
#puts "EQ_REL_COMP: #{set1} #{op} #{set2}"
|
||||
case b
|
||||
when 'true', 'false'
|
||||
b = Functions::boolean( b )
|
||||
for v in a
|
||||
v = Functions::boolean(v)
|
||||
return true if compare( v, op, b )
|
||||
end
|
||||
when /^\d+(\.\d+)?$/
|
||||
b = Functions::number( b )
|
||||
#puts "B = #{b.inspect}"
|
||||
for v in a
|
||||
#puts "v = #{v.inspect}"
|
||||
v = Functions::number(v)
|
||||
#puts "v = #{v.inspect}"
|
||||
#puts compare(v,op,b)
|
||||
return true if compare( v, op, b )
|
||||
end
|
||||
else
|
||||
b = Functions::string( b )
|
||||
for v in a
|
||||
v = Functions::string(v)
|
||||
return true if compare( v, op, b )
|
||||
end
|
||||
end
|
||||
else
|
||||
# If neither is nodeset,
|
||||
# If op is = or !=
|
||||
# If either boolean, convert to boolean
|
||||
# If either number, convert to number
|
||||
# Else, convert to string
|
||||
# Else
|
||||
# Convert both to numbers and compare
|
||||
s1 = set1.to_s
|
||||
s2 = set2.to_s
|
||||
#puts "EQ_REL_COMP: #{set1}=>#{s1}, #{set2}=>#{s2}"
|
||||
if s1 == 'true' or s1 == 'false' or s2 == 'true' or s2 == 'false'
|
||||
#puts "Functions::boolean(#{set1})=>#{Functions::boolean(set1)}"
|
||||
#puts "Functions::boolean(#{set2})=>#{Functions::boolean(set2)}"
|
||||
set1 = Functions::boolean( set1 )
|
||||
set2 = Functions::boolean( set2 )
|
||||
else
|
||||
if op == :eq or op == :neq
|
||||
if s1 =~ /^\d+(\.\d+)?$/ or s2 =~ /^\d+(\.\d+)?$/
|
||||
set1 = Functions::number( s1 )
|
||||
set2 = Functions::number( s2 )
|
||||
else
|
||||
set1 = Functions::string( set1 )
|
||||
set2 = Functions::string( set2 )
|
||||
end
|
||||
else
|
||||
set1 = Functions::number( set1 )
|
||||
set2 = Functions::number( set2 )
|
||||
end
|
||||
end
|
||||
#puts "EQ_REL_COMP: #{set1} #{op} #{set2}"
|
||||
#puts ">>> #{compare( set1, op, set2 )}"
|
||||
return compare( set1, op, set2 )
|
||||
end
|
||||
return false
|
||||
end
|
||||
return compare( set1, op, set2 )
|
||||
end
|
||||
return false
|
||||
end
|
||||
|
||||
def compare a, op, b
|
||||
def compare a, op, b
|
||||
#puts "COMPARE #{a.to_s} #{op} #{b.to_s}"
|
||||
case op
|
||||
when :eq
|
||||
a == b
|
||||
when :neq
|
||||
a != b
|
||||
when :lt
|
||||
a < b
|
||||
when :lteq
|
||||
a <= b
|
||||
when :gt
|
||||
a > b
|
||||
when :gteq
|
||||
a >= b
|
||||
when :and
|
||||
a and b
|
||||
when :or
|
||||
a or b
|
||||
else
|
||||
false
|
||||
end
|
||||
end
|
||||
end
|
||||
case op
|
||||
when :eq
|
||||
a == b
|
||||
when :neq
|
||||
a != b
|
||||
when :lt
|
||||
a < b
|
||||
when :lteq
|
||||
a <= b
|
||||
when :gt
|
||||
a > b
|
||||
when :gteq
|
||||
a >= b
|
||||
when :and
|
||||
a and b
|
||||
when :or
|
||||
a or b
|
||||
else
|
||||
false
|
||||
end
|
||||
end
|
||||
end
|
||||
end
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue