/preinstalled-gems/data/gems/rdoc-2.5.1/lib/rdoc/parser/c.rb
Ruby | 708 lines | 453 code | 90 blank | 165 comment | 49 complexity | 3d640ee715f43e04f4c6b5c408d32827 MD5 | raw file
Possible License(s): BSD-3-Clause, MPL-2.0-no-copyleft-exception, 0BSD, GPL-2.0, LGPL-2.1
- require 'rdoc/parser'
- require 'rdoc/parser/ruby'
- require 'rdoc/known_classes'
- ##
- # We attempt to parse C extension files. Basically we look for
- # the standard patterns that you find in extensions: <tt>rb_define_class,
- # rb_define_method</tt> and so on. We also try to find the corresponding
- # C source for the methods and extract comments, but if we fail
- # we don't worry too much.
- #
- # The comments associated with a Ruby method are extracted from the C
- # comment block associated with the routine that _implements_ that
- # method, that is to say the method whose name is given in the
- # <tt>rb_define_method</tt> call. For example, you might write:
- #
- # /*
- # * Returns a new array that is a one-dimensional flattening of this
- # * array (recursively). That is, for every element that is an array,
- # * extract its elements into the new array.
- # *
- # * s = [ 1, 2, 3 ] #=> [1, 2, 3]
- # * t = [ 4, 5, 6, [7, 8] ] #=> [4, 5, 6, [7, 8]]
- # * a = [ s, t, 9, 10 ] #=> [[1, 2, 3], [4, 5, 6, [7, 8]], 9, 10]
- # * a.flatten #=> [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
- # */
- # static VALUE
- # rb_ary_flatten(ary)
- # VALUE ary;
- # {
- # ary = rb_obj_dup(ary);
- # rb_ary_flatten_bang(ary);
- # return ary;
- # }
- #
- # ...
- #
- # void
- # Init_Array()
- # {
- # ...
- # rb_define_method(rb_cArray, "flatten", rb_ary_flatten, 0);
- #
- # Here RDoc will determine from the rb_define_method line that there's a
- # method called "flatten" in class Array, and will look for the implementation
- # in the method rb_ary_flatten. It will then use the comment from that
- # method in the HTML output. This method must be in the same source file
- # as the rb_define_method.
- #
- # The comment blocks may include special directives:
- #
- # [Document-class: <i>name</i>]
- # This comment block is documentation for the given class. Use this
- # when the <tt>Init_xxx</tt> method is not named after the class.
- #
- # [Document-method: <i>name</i>]
- # This comment documents the named method. Use when RDoc cannot
- # automatically find the method from it's declaration
- #
- # [call-seq: <i>text up to an empty line</i>]
- # Because C source doesn't give descripive names to Ruby-level parameters,
- # you need to document the calling sequence explicitly
- #
- # In addition, RDoc assumes by default that the C method implementing a
- # Ruby function is in the same source file as the rb_define_method call.
- # If this isn't the case, add the comment:
- #
- # rb_define_method(....); // in filename
- #
- # As an example, we might have an extension that defines multiple classes
- # in its Init_xxx method. We could document them using
- #
- # /*
- # * Document-class: MyClass
- # *
- # * Encapsulate the writing and reading of the configuration
- # * file. ...
- # */
- #
- # /*
- # * Document-method: read_value
- # *
- # * call-seq:
- # * cfg.read_value(key) -> value
- # * cfg.read_value(key} { |key| } -> value
- # *
- # * Return the value corresponding to +key+ from the configuration.
- # * In the second form, if the key isn't found, invoke the
- # * block and return its value.
- # */
- class RDoc::Parser::C < RDoc::Parser
- parse_files_matching(/\.(?:([CcHh])\1?|c([+xp])\2|y)\z/)
- include RDoc::Text
- ##
- # C file the parser is parsing
- attr_accessor :content
- ##
- # Resets cross-file state. Call when parsing different projects that need
- # separate documentation.
- def self.reset
- @@enclosure_classes = {}
- @@known_bodies = {}
- end
- reset
- ##
- # Prepare to parse a C file
- def initialize(top_level, file_name, content, options, stats)
- super
- @known_classes = RDoc::KNOWN_CLASSES.dup
- @content = handle_tab_width handle_ifdefs_in(@content)
- @classes = Hash.new
- @file_dir = File.dirname(@file_name)
- end
- def do_aliases
- @content.scan(%r{rb_define_alias\s*\(\s*(\w+),\s*"([^"]+)",\s*"([^"]+)"\s*\)}m) do
- |var_name, new_name, old_name|
- class_name = @known_classes[var_name] || var_name
- class_obj = find_class(var_name, class_name)
- as = class_obj.add_alias RDoc::Alias.new("", old_name, new_name, "")
- @stats.add_alias as
- end
- end
- def do_classes
- @content.scan(/(\w+)\s* = \s*rb_define_module\s*\(\s*"(\w+)"\s*\)/mx) do
- |var_name, class_name|
- handle_class_module(var_name, "module", class_name, nil, nil)
- end
- # The '.' lets us handle SWIG-generated files
- @content.scan(/([\w\.]+)\s* = \s*rb_define_class\s*
- \(
- \s*"(\w+)",
- \s*(\w+)\s*
- \)/mx) do |var_name, class_name, parent|
- handle_class_module(var_name, "class", class_name, parent, nil)
- end
- @content.scan(/(\w+)\s*=\s*boot_defclass\s*\(\s*"(\w+?)",\s*(\w+?)\s*\)/) do
- |var_name, class_name, parent|
- parent = nil if parent == "0"
- handle_class_module(var_name, "class", class_name, parent, nil)
- end
- @content.scan(/(\w+)\s* = \s*rb_define_module_under\s*
- \(
- \s*(\w+),
- \s*"(\w+)"
- \s*\)/mx) do |var_name, in_module, class_name|
- handle_class_module(var_name, "module", class_name, nil, in_module)
- end
- @content.scan(/([\w\.]+)\s* = \s*rb_define_class_under\s*
- \(
- \s*(\w+),
- \s*"(\w+)",
- \s*([\w\*\s\(\)\.\->]+)\s* # for SWIG
- \s*\)/mx) do |var_name, in_module, class_name, parent|
- handle_class_module(var_name, "class", class_name, parent, in_module)
- end
- end
- def do_constants
- @content.scan(%r{\Wrb_define_
- ( variable |
- readonly_variable |
- const |
- global_const | )
- \s*\(
- (?:\s*(\w+),)?
- \s*"(\w+)",
- \s*(.*?)\s*\)\s*;
- }xm) do |type, var_name, const_name, definition|
- var_name = "rb_cObject" if !var_name or var_name == "rb_mKernel"
- handle_constants type, var_name, const_name, definition
- end
- end
- ##
- # Look for includes of the form:
- #
- # rb_include_module(rb_cArray, rb_mEnumerable);
- def do_includes
- @content.scan(/rb_include_module\s*\(\s*(\w+?),\s*(\w+?)\s*\)/) do |c,m|
- if cls = @classes[c]
- m = @known_classes[m] || m
- cls.add_include RDoc::Include.new(m, "")
- end
- end
- end
- def do_methods
- @content.scan(%r{rb_define_
- (
- singleton_method |
- method |
- module_function |
- private_method
- )
- \s*\(\s*([\w\.]+),
- \s*"([^"]+)",
- \s*(?:RUBY_METHOD_FUNC\(|VALUEFUNC\()?(\w+)\)?,
- \s*(-?\w+)\s*\)
- (?:;\s*/[*/]\s+in\s+(\w+?\.[cy]))?
- }xm) do
- |type, var_name, meth_name, meth_body, param_count, source_file|
- # Ignore top-object and weird struct.c dynamic stuff
- next if var_name == "ruby_top_self"
- next if var_name == "nstr"
- next if var_name == "envtbl"
- next if var_name == "argf" # it'd be nice to handle this one
- var_name = "rb_cObject" if var_name == "rb_mKernel"
- handle_method(type, var_name, meth_name, meth_body, param_count,
- source_file)
- end
- @content.scan(%r{rb_define_attr\(
- \s*([\w\.]+),
- \s*"([^"]+)",
- \s*(\d+),
- \s*(\d+)\s*\);
- }xm) do |var_name, attr_name, attr_reader, attr_writer|
- #var_name = "rb_cObject" if var_name == "rb_mKernel"
- handle_attr(var_name, attr_name,
- attr_reader.to_i != 0,
- attr_writer.to_i != 0)
- end
- @content.scan(%r{rb_define_global_function\s*\(
- \s*"([^"]+)",
- \s*(?:RUBY_METHOD_FUNC\(|VALUEFUNC\()?(\w+)\)?,
- \s*(-?\w+)\s*\)
- (?:;\s*/[*/]\s+in\s+(\w+?\.[cy]))?
- }xm) do |meth_name, meth_body, param_count, source_file|
- handle_method("method", "rb_mKernel", meth_name,
- meth_body, param_count, source_file)
- end
- @content.scan(/define_filetest_function\s*\(
- \s*"([^"]+)",
- \s*(?:RUBY_METHOD_FUNC\(|VALUEFUNC\()?(\w+)\)?,
- \s*(-?\w+)\s*\)/xm) do
- |meth_name, meth_body, param_count|
- handle_method("method", "rb_mFileTest", meth_name, meth_body, param_count)
- handle_method("singleton_method", "rb_cFile", meth_name, meth_body, param_count)
- end
- end
- def find_attr_comment(attr_name)
- if @content =~ %r{((?>/\*.*?\*/\s+))
- rb_define_attr\((?:\s*(\w+),)?\s*"#{attr_name}"\s*,.*?\)\s*;}xmi
- $1
- elsif @content =~ %r{Document-attr:\s#{attr_name}\s*?\n((?>.*?\*/))}m
- $1
- else
- ''
- end
- end
- ##
- # Find the C code corresponding to a Ruby method
- def find_body(class_name, meth_name, meth_obj, body, quiet = false)
- case body
- when %r"((?>/\*.*?\*/\s*))((?:(?:static|SWIGINTERN)\s+)?(?:intern\s+)?VALUE\s+#{meth_name}
- \s*(\([^)]*\))([^;]|$))"xm
- comment = $1
- body_text = $2
- params = $3
- remove_private_comments comment if comment
- # see if we can find the whole body
- re = Regexp.escape(body_text) + '[^(]*^\{.*?^\}'
- body_text = $& if /#{re}/m =~ body
- # The comment block may have been overridden with a 'Document-method'
- # block. This happens in the interpreter when multiple methods are
- # vectored through to the same C method but those methods are logically
- # distinct (for example Kernel.hash and Kernel.object_id share the same
- # implementation
- override_comment = find_override_comment class_name, meth_obj.name
- comment = override_comment if override_comment
- find_modifiers comment, meth_obj if comment
- #meth_obj.params = params
- meth_obj.start_collecting_tokens
- tk = RDoc::RubyToken::Token.new nil, 1, 1
- tk.set_text body_text
- meth_obj.add_token tk
- meth_obj.comment = strip_stars comment
- when %r{((?>/\*.*?\*/\s*))^\s*(\#\s*define\s+#{meth_name}\s+(\w+))}m
- comment = $1
- body_text = $2
- find_body class_name, $3, meth_obj, body, true
- find_modifiers comment, meth_obj
- meth_obj.start_collecting_tokens
- tk = RDoc::RubyToken::Token.new nil, 1, 1
- tk.set_text body_text
- meth_obj.add_token tk
- meth_obj.comment = strip_stars(comment) + meth_obj.comment.to_s
- when %r{^\s*\#\s*define\s+#{meth_name}\s+(\w+)}m
- unless find_body(class_name, $1, meth_obj, body, true)
- warn "No definition for #{meth_name}" unless @options.quiet
- return false
- end
- else
- # No body, but might still have an override comment
- comment = find_override_comment(class_name, meth_obj.name)
- if comment
- find_modifiers(comment, meth_obj)
- meth_obj.comment = strip_stars comment
- else
- warn "No definition for #{meth_name}" unless @options.quiet
- return false
- end
- end
- true
- end
- def find_class(raw_name, name)
- unless @classes[raw_name]
- if raw_name =~ /^rb_m/
- container = @top_level.add_module RDoc::NormalModule, name
- else
- container = @top_level.add_class RDoc::NormalClass, name
- end
- container.record_location @top_level
- @classes[raw_name] = container
- end
- @classes[raw_name]
- end
- ##
- # Look for class or module documentation above Init_+class_name+(void),
- # in a Document-class +class_name+ (or module) comment or above an
- # rb_define_class (or module). If a comment is supplied above a matching
- # Init_ and a rb_define_class the Init_ comment is used.
- #
- # /*
- # * This is a comment for Foo
- # */
- # Init_Foo(void) {
- # VALUE cFoo = rb_define_class("Foo", rb_cObject);
- # }
- #
- # /*
- # * Document-class: Foo
- # * This is a comment for Foo
- # */
- # Init_foo(void) {
- # VALUE cFoo = rb_define_class("Foo", rb_cObject);
- # }
- #
- # /*
- # * This is a comment for Foo
- # */
- # VALUE cFoo = rb_define_class("Foo", rb_cObject);
- def find_class_comment(class_name, class_mod)
- comment = nil
- if @content =~ %r{
- ((?>/\*.*?\*/\s+))
- (static\s+)?
- void\s+
- Init_#{class_name}\s*(?:_\(\s*)?\(\s*(?:void\s*)?\)}xmi then # )
- comment = $1
- elsif @content =~ %r{Document-(?:class|module):\s+#{class_name}\s*?(?:<\s+[:,\w]+)?\n((?>.*?\*/))}m then # "
- comment = $1
- elsif @content =~ %r{((?>/\*.*?\*/\s+))
- ([\w\.\s]+\s* = \s+)?rb_define_(class|module).*?"(#{class_name})"}xm then
- comment = $1
- end
- return unless comment
- comment = strip_stars comment
- comment = look_for_directives_in class_mod, comment
- class_mod.comment = comment
- end
- ##
- # Finds a comment matching +type+ and +const_name+ either above the
- # comment or in the matching Document- section.
- def find_const_comment(type, const_name)
- if @content =~ %r{((?>^\s*/\*.*?\*/\s+))
- rb_define_#{type}\((?:\s*(\w+),)?\s*"#{const_name}"\s*,.*?\)\s*;}xmi
- $1
- elsif @content =~ %r{Document-(?:const|global|variable):\s#{const_name}\s*?\n((?>.*?\*/))}m
- $1
- else
- ''
- end
- end
- ##
- # If the comment block contains a section that looks like:
- #
- # call-seq:
- # Array.new
- # Array.new(10)
- #
- # use it for the parameters.
- def find_modifiers(comment, meth_obj)
- if comment.sub!(/:nodoc:\s*^\s*\*?\s*$/m, '') or
- comment.sub!(/\A\/\*\s*:nodoc:\s*\*\/\Z/, '')
- meth_obj.document_self = false
- end
- if comment.sub!(/call-seq:(.*?)^\s*\*?\s*$/m, '') or
- comment.sub!(/\A\/\*\s*call-seq:(.*?)\*\/\Z/, '')
- seq = $1
- seq.gsub!(/^\s*\*\s*/, '')
- meth_obj.call_seq = seq
- end
- end
- def find_override_comment(class_name, meth_name)
- name = Regexp.escape(meth_name)
- if @content =~ %r{Document-method:\s+#{class_name}(?:\.|::|#)#{name}\s*?\n((?>.*?\*/))}m then
- $1
- elsif @content =~ %r{Document-method:\s#{name}\s*?\n((?>.*?\*/))}m then
- $1
- end
- end
- def handle_attr(var_name, attr_name, reader, writer)
- rw = ''
- rw << 'R' if reader
- rw << 'W' if writer
- class_name = @known_classes[var_name]
- return unless class_name
- class_obj = find_class(var_name, class_name)
- if class_obj
- comment = find_attr_comment(attr_name)
- comment = strip_stars comment
- att = RDoc::Attr.new '', attr_name, rw, comment
- @stats.add_method att
- class_obj.add_attribute(att)
- end
- end
- def handle_class_module(var_name, type, class_name, parent, in_module)
- parent_name = @known_classes[parent] || parent
- if in_module then
- enclosure = @classes[in_module] || @@enclosure_classes[in_module]
- if enclosure.nil? and enclosure = @known_classes[in_module] then
- type = /^rb_m/ =~ in_module ? "module" : "class"
- handle_class_module in_module, type, enclosure, nil, nil
- enclosure = @classes[in_module]
- end
- unless enclosure then
- warn "Enclosing class/module '#{in_module}' for #{type} #{class_name} not known"
- return
- end
- else
- enclosure = @top_level
- end
- if type == "class" then
- full_name = if RDoc::ClassModule === enclosure then
- enclosure.full_name + "::#{class_name}"
- else
- class_name
- end
- if @content =~ %r{Document-class:\s+#{full_name}\s*<\s+([:,\w]+)} then
- parent_name = $1
- end
- cm = enclosure.add_class RDoc::NormalClass, class_name, parent_name
- @stats.add_class cm
- else
- cm = enclosure.add_module RDoc::NormalModule, class_name
- @stats.add_module cm
- end
- cm.record_location enclosure.top_level
- find_class_comment cm.full_name, cm
- @classes[var_name] = cm
- @@enclosure_classes[var_name] = cm
- @known_classes[var_name] = cm.full_name
- end
- ##
- # Adds constant comments. By providing some_value: at the start ofthe
- # comment you can override the C value of the comment to give a friendly
- # definition.
- #
- # /* 300: The perfect score in bowling */
- # rb_define_const(cFoo, "PERFECT", INT2FIX(300);
- #
- # Will override +INT2FIX(300)+ with the value +300+ in the output RDoc.
- # Values may include quotes and escaped colons (\:).
- def handle_constants(type, var_name, const_name, definition)
- class_name = @known_classes[var_name]
- return unless class_name
- class_obj = find_class var_name, class_name
- unless class_obj then
- warn "Enclosing class/module #{const_name.inspect} not known"
- return
- end
- comment = find_const_comment type, const_name
- comment = strip_stars comment
- comment = normalize_comment comment
- # In the case of rb_define_const, the definition and comment are in
- # "/* definition: comment */" form. The literal ':' and '\' characters
- # can be escaped with a backslash.
- if type.downcase == 'const' then
- elements = comment.split ':'
- if elements.nil? or elements.empty? then
- con = RDoc::Constant.new const_name, definition, comment
- else
- new_definition = elements[0..-2].join(':')
- if new_definition.empty? then # Default to literal C definition
- new_definition = definition
- else
- new_definition.gsub!("\:", ":")
- new_definition.gsub!("\\", '\\')
- end
- new_definition.sub!(/\A(\s+)/, '')
- new_comment = if $1.nil? then
- elements.last.lstrip
- else
- "#{$1}#{elements.last.lstrip}"
- end
- con = RDoc::Constant.new const_name, new_definition, new_comment
- end
- else
- con = RDoc::Constant.new const_name, definition, comment
- end
- @stats.add_constant con
- class_obj.add_constant con
- end
- ##
- # Removes #ifdefs that would otherwise confuse us
- def handle_ifdefs_in(body)
- body.gsub(/^#ifdef HAVE_PROTOTYPES.*?#else.*?\n(.*?)#endif.*?\n/m, '\1')
- end
- def handle_method(type, var_name, meth_name, meth_body, param_count,
- source_file = nil)
- class_name = @known_classes[var_name]
- return unless class_name
- class_obj = find_class var_name, class_name
- if class_obj then
- if meth_name == "initialize" then
- meth_name = "new"
- type = "singleton_method"
- end
- meth_obj = RDoc::AnyMethod.new '', meth_name
- meth_obj.singleton = %w[singleton_method module_function].include? type
- p_count = Integer(param_count) rescue -1
- if p_count < 0 then
- meth_obj.params = "(...)"
- elsif p_count == 0
- meth_obj.params = "()"
- else
- meth_obj.params = "(" + (1..p_count).map{|i| "p#{i}"}.join(", ") + ")"
- end
- if source_file then
- file_name = File.join @file_dir, source_file
- if File.exist? file_name then
- body = (@@known_bodies[file_name] ||= File.read(file_name))
- else
- warn "unknown source #{source_file} for #{meth_name} in #{@file_name}"
- end
- else
- body = @content
- end
- if find_body(class_name, meth_body, meth_obj, body) and meth_obj.document_self then
- class_obj.add_method meth_obj
- @stats.add_method meth_obj
- meth_obj.visibility = :private if 'private_method' == type
- end
- end
- end
- def handle_tab_width(body)
- if /\t/ =~ body
- tab_width = @options.tab_width
- body.split(/\n/).map do |line|
- 1 while line.gsub!(/\t+/) { ' ' * (tab_width*$&.length - $`.length % tab_width)} && $~ #`
- line
- end .join("\n")
- else
- body
- end
- end
- ##
- # Look for directives in a normal comment block:
- #
- # /*
- # * :title: My Awesome Project
- # */
- #
- # This routine modifies it's parameter
- def look_for_directives_in(context, comment)
- preprocess = RDoc::Markup::PreProcess.new @file_name, @options.rdoc_include
- preprocess.handle comment do |directive, param|
- case directive
- when 'main' then
- @options.main_page = param
- ''
- when 'title' then
- @options.title = param
- ''
- else
- warn "Unrecognized directive :#{directive}:"
- false
- end
- end
- comment
- end
- ##
- # Removes lines that are commented out that might otherwise get picked up
- # when scanning for classes and methods
- def remove_commented_out_lines
- @content.gsub!(%r{//.*rb_define_}, '//')
- end
- def remove_private_comments(comment)
- comment.gsub!(/\/?\*--\n(.*?)\/?\*\+\+/m, '')
- comment.sub!(/\/?\*--\n.*/m, '')
- end
- ##
- # Extract the classes/modules and methods from a C file and return the
- # corresponding top-level object
- def scan
- remove_commented_out_lines
- do_classes
- do_constants
- do_methods
- do_includes
- do_aliases
- @top_level
- end
- end