lib/rdoc/rd/block_parser.ry
class BlockParser
preclow
nonassoc DUMMY
left ITEMLISTLINE
ENUMLISTLINE
DESCLISTLINE
METHODLISTLINE
STRINGLINE
prechigh
token STRINGLINE
ITEMLISTLINE
ENUMLISTLINE
DESCLISTLINE
METHODLISTLINE
WHITELINE
SUBTREE
HEADLINE
INCLUDE
INDENT
DEDENT
DUMMY
rule
document : blocks { result = RDoc::Markup::Document.new(*val[0]) }
| { raise ParseError, "file empty" }
;
blocks : blocks block { result = val[0].concat val[1] }
| block { result = val[0] }
;
block : textblock { result = val }
| verbatim { result = val }
| lists
| headline { result = val }
| include { result = val }
| WHITELINE { result = [RDoc::Markup::BlankLine.new] }
| SUBTREE { result = val[0].parts }
;
headline : HEADLINE {
# val[0] is like [level, title]
title = @inline_parser.parse(val[0][1])
result = RDoc::Markup::Heading.new(val[0][0], title)
}
;
include : INCLUDE {
result = RDoc::Markup::Include.new val[0], @include_path
}
;
textblock : textblockcontent = DUMMY {
# val[0] is Array of String
result = paragraph val[0]
}
;
textblockcontent : textblockcontent STRINGLINE { result << val[1].rstrip }
| STRINGLINE { result = [val[0].rstrip] }
;
verbatim : INDENT verbatimcontent DEDENT {
# val[1] is Array of String
content = cut_off val[1]
result = RDoc::Markup::Verbatim.new(*content)
# imform to lexer.
@in_verbatim = false
}
;
verbatim_after_lists : verbatimcontent {
# val[0] is Array of String
content = cut_off val[0]
result = RDoc::Markup::Verbatim.new(*content)
# imform to lexer.
@in_verbatim = false
}
;
verbatimcontent : verbatimcontent STRINGLINE {
result << val[1]
} | verbatimcontent INDENT verbatimcontent DEDENT {
result.concat val[2]
} | verbatimcontent WHITELINE {
result << "\n"
} | STRINGLINE {
result = val
# inform to lexer.
@in_verbatim = true
}
;
list : itemlist
| enumlist
| desclist
| methodlist
;
lists : lists2 = DUMMY {
result = val[0]
} | INDENT lists2 DEDENT {
result = val[1]
} | INDENT lists2 verbatim_after_lists DEDENT {
result = val[1].push(val[2])
}
;
lists2 : lists2 list { result = val[0] << val[1] }
| list { result = [val[0]] }
;
itemlist : itemlistitems = DUMMY {
result = RDoc::Markup::List.new :BULLET, *val[0]
}
;
itemlistitems : itemlistitems itemlistitem { result.push(val[1]) }
| itemlistitem { result = val }
;
itemlistitem : first_textblock_in_itemlist other_blocks_in_list DEDENT {
result = RDoc::Markup::ListItem.new nil, val[0], *val[1]
}
;
enumlist : enumlistitems = DUMMY {
result = RDoc::Markup::List.new :NUMBER, *val[0]
}
;
enumlistitems : enumlistitems enumlistitem { result.push(val[1]) }
| enumlistitem { result = val }
;
enumlistitem : first_textblock_in_enumlist other_blocks_in_list DEDENT {
result = RDoc::Markup::ListItem.new nil, val[0], *val[1]
}
;
desclist : desclistitems = DUMMY {
result = RDoc::Markup::List.new :NOTE, *val[0]
}
;
desclistitems : desclistitems desclistitem { result.push(val[1]) }
| desclistitem { result = val }
;
desclistitem : DESCLISTLINE description_part DEDENT {
term = @inline_parser.parse val[0].strip
result = RDoc::Markup::ListItem.new term, *val[1]
}
;
methodlist : methodlistitems = DUMMY {
result = RDoc::Markup::List.new :LABEL, *val[0]
}
;
methodlistitems : methodlistitems methodlistitem { result.push(val[1]) }
| methodlistitem { result = val }
;
methodlistitem : METHODLISTLINE description_part DEDENT {
result = RDoc::Markup::ListItem.new "<tt>#{val[0].strip}</tt>", *val[1]
}
;
description_part : whitelines textblock blocks_in_list {
result = [val[1]].concat(val[2])
} | whitelines textblock {
result = [val[1]]
} | whitelines INDENT blocks_in_list DEDENT {
result = val[2]
} | whitelines {
result = []
}
;
blocks_in_list : blocks_in_list block_in_list { result.concat val[1] }
| block_in_list
;
block_in_list : textblock { result = val }
| verbatim { result = val }
| lists
| WHITELINE { result = [] }
;
whitelines : whitelines2
|
;
whitelines2 : WHITELINE whitelines2
| WHITELINE
;
first_textblock_in_itemlist : ITEMLISTLINE textblockcontent {
result = paragraph [val[0]].concat(val[1])
} | ITEMLISTLINE {
result = paragraph [val[0]]
}
;
first_textblock_in_enumlist : ENUMLISTLINE textblockcontent {
result = paragraph [val[0]].concat(val[1])
} | ENUMLISTLINE {
result = paragraph [val[0]]
}
;
other_blocks_in_list : verbatim blocks_in_list {
result = [val[0]].concat(val[1])
} | lists blocks_in_list { result.concat val[1] }
| WHITELINE blocks_in_list { result = val[1] }
| verbatim { result = val }
| lists
| WHITELINE { result = [] }
| { result = [] }
;
end
---- inner
# :stopdoc:
MARK_TO_LEVEL = {
'=' => 1,
'==' => 2,
'===' => 3,
'====' => 4,
'+' => 5,
'++' => 6,
}
# :startdoc:
##
# Footnotes for this document
attr_reader :footnotes
##
# Labels for items in this document
attr_reader :labels
##
# Path to find included files in
attr_accessor :include_path
##
# Creates a new RDoc::RD::BlockParser. Use #parse to parse an rd-format
# document.
def initialize
@inline_parser = RDoc::RD::InlineParser.new self
@include_path = []
# for testing
@footnotes = []
@labels = {}
end
##
# Parses +src+ and returns an RDoc::Markup::Document.
def parse src
@src = src
@src.push false
@footnotes = []
@labels = {}
# @i: index(line no.) of src
@i = 0
# stack for current indentation
@indent_stack = []
# how indented.
@current_indent = @indent_stack.join("")
# RDoc::RD::BlockParser for tmp src
@subparser = nil
# which part is in now
@in_part = nil
@part_content = []
@in_verbatim = false
@yydebug = true
document = do_parse
unless @footnotes.empty? then
blankline = document.parts.pop
document.parts << RDoc::Markup::Rule.new(1)
document.parts.concat @footnotes
document.parts.push blankline
end
document
end
##
# Returns the next token from the document
def next_token # :nodoc:
# preprocessing
# if it is not in RD part
# => method
while @in_part != "rd"
line = @src[@i]
@i += 1 # next line
case line
# src end
when false
return [false, false]
# RD part begin
when /^=begin\s*(?:\bRD\b.*)?\s*$/
if @in_part # if in non-RD part
@part_content.push(line)
else
@in_part = "rd"
return [:WHITELINE, "=begin\n"] # <= for textblockand
end
# non-RD part begin
when /^=begin\s+(\w+)/
part = $1
=begin # not imported to RDoc
if @in_part # if in non-RD part
@part_content.push(line)
else
@in_part = part if @tree.filter[part] # if filter exists
# p "BEGIN_PART: #{@in_part}" # DEBUG
end
=end
@in_part = part
# non-RD part end
when /^=end(?:$|[\s\0\C-d\C-z])/
if @in_part # if in non-RD part
=begin # not imported to RDoc
# p "END_PART: #{@in_part}" # DEBUG
# make Part-in object
part = RDoc::RD::Part.new(@part_content.join(""), @tree, "r")
@part_content.clear
# call filter, part_out is output(Part object)
part_out = @tree.filter[@in_part].call(part)
if @tree.filter[@in_part].mode == :rd # if output is RD formatted
subtree = parse_subtree(part_out.to_a)
else # if output is target formatted
basename = Tempfile.create(["rdtmp", ".#{@in_part}"], @tree.tmp_dir) do |tmpfile|
tmpfile.print(part_out)
File.basename(tmpfile.path)
end
subtree = parse_subtree(["=begin\n", "<<< #{basename}\n", "=end\n"])
end
@in_part = nil
return [:SUBTREE, subtree]
=end
end
else
=begin # not imported to RDoc
if @in_part # if in non-RD part
@part_content.push(line)
end
=end
end
end
@current_indent = @indent_stack.join("")
line = @src[@i]
case line
when false
if_current_indent_equal("") do
[false, false]
end
when /^=end/
if_current_indent_equal("") do
@in_part = nil
[:WHITELINE, "=end"] # MUST CHANGE??
end
when /^\s*$/
@i += 1 # next line
return [:WHITELINE, ':WHITELINE']
when /^\#/ # comment line
@i += 1 # next line
self.next_token()
when /^(={1,4})(?!=)\s*(?=\S)/, /^(\+{1,2})(?!\+)\s*(?=\S)/
rest = $' # '
rest.strip!
mark = $1
if_current_indent_equal("") do
return [:HEADLINE, [MARK_TO_LEVEL[mark], rest]]
end
when /^<<<\s*(\S+)/
file = $1
if_current_indent_equal("") do
suffix = file[-3 .. -1]
if suffix == ".rd" or suffix == ".rb"
subtree = parse_subtree(get_included(file))
[:SUBTREE, subtree]
else
[:INCLUDE, file]
end
end
when /^(\s*)\*(\s*)/
rest = $' # '
newIndent = $2
if_current_indent_equal($1) do
if @in_verbatim
[:STRINGLINE, line]
else
@indent_stack.push("\s" + newIndent)
[:ITEMLISTLINE, rest]
end
end
when /^(\s*)(\(\d+\))(\s*)/
rest = $' # '
mark = $2
newIndent = $3
if_current_indent_equal($1) do
if @in_verbatim
[:STRINGLINE, line]
else
@indent_stack.push("\s" * mark.size + newIndent)
[:ENUMLISTLINE, rest]
end
end
when /^(\s*):(\s*)/
rest = $' # '
newIndent = $2
if_current_indent_equal($1) do
if @in_verbatim
[:STRINGLINE, line]
else
@indent_stack.push("\s#{$2}")
[:DESCLISTLINE, rest]
end
end
when /^(\s*)---(?!-|\s*$)/
indent = $1
rest = $'
/\s*/ === rest
term = $'
new_indent = $&
if_current_indent_equal(indent) do
if @in_verbatim
[:STRINGLINE, line]
else
@indent_stack.push("\s\s\s" + new_indent)
[:METHODLISTLINE, term]
end
end
when /^(\s*)/
if_current_indent_equal($1) do
[:STRINGLINE, line]
end
else
raise "[BUG] parsing error may occurred."
end
end
##
# Yields to the given block if +indent+ matches the current indent, otherwise
# an indentation token is processed.
def if_current_indent_equal(indent)
indent = indent.sub(/\t/, "\s" * 8)
if @current_indent == indent
@i += 1 # next line
yield
elsif indent.index(@current_indent) == 0
@indent_stack.push(indent[@current_indent.size .. -1])
[:INDENT, ":INDENT"]
else
@indent_stack.pop
[:DEDENT, ":DEDENT"]
end
end
private :if_current_indent_equal
##
# Cuts off excess whitespace in +src+
def cut_off(src)
ret = []
whiteline_buf = []
line = src.shift
/^\s*/ =~ line
indent = Regexp.quote($&)
ret.push($')
while line = src.shift
if /^(\s*)$/ =~ line
whiteline_buf.push(line)
elsif /^#{indent}/ =~ line
unless whiteline_buf.empty?
ret.concat(whiteline_buf)
whiteline_buf.clear
end
ret.push($')
else
raise "[BUG]: probably Parser Error while cutting off.\n"
end
end
ret
end
private :cut_off
def set_term_to_element(parent, term)
# parent.set_term_under_document_struct(term, @tree.document_struct)
parent.set_term_without_document_struct(term)
end
private :set_term_to_element
##
# Raises a ParseError when invalid formatting is found
def on_error(et, ev, _values)
prv, cur, nxt = format_line_num(@i, @i+1, @i+2)
raise ParseError, <<Msg
RD syntax error: line #{@i+1}:
#{prv} |#{@src[@i-1].chomp}
#{cur}=>|#{@src[@i].chomp}
#{nxt} |#{@src[@i+1].chomp}
Msg
end
##
# Current line number
def line_index
@i
end
##
# Parses subtree +src+
def parse_subtree src
@subparser ||= RDoc::RD::BlockParser.new
@subparser.parse src
end
private :parse_subtree
##
# Retrieves the content for +file+ from the include_path
def get_included(file)
included = []
@include_path.each do |dir|
file_name = File.join dir, file
if File.exist? file_name then
included = File.readlines file_name
break
end
end
included
end
private :get_included
##
# Formats line numbers +line_numbers+ prettily
def format_line_num(*line_numbers)
width = line_numbers.collect{|i| i.to_s.length }.max
line_numbers.collect{|i| sprintf("%#{width}d", i) }
end
private :format_line_num
##
# Retrieves the content of +values+ as a single String
def content values
values.map { |value| value.content }.join
end
##
# Creates a paragraph for +value+
def paragraph value
content = cut_off(value).join(' ').rstrip
contents = @inline_parser.parse content
RDoc::Markup::Paragraph.new(*contents)
end
##
# Adds footnote +content+ to the document
def add_footnote content
index = @footnotes.length / 2 + 1
footmark_link = "{^#{index}}[rdoc-label:footmark-#{index}:foottext-#{index}]"
@footnotes << RDoc::Markup::Paragraph.new(footmark_link, ' ', *content)
@footnotes << RDoc::Markup::BlankLine.new
index
end
##
# Adds label +label+ to the document
def add_label label
@labels[label] = true
label
end
# :stopdoc:
---- header
class RDoc::RD
##
# RD format parser for headings, paragraphs, lists, verbatim sections that
# exist as blocks.
---- footer
end