# $Id$ # == Usage # # saikuro [ -h ] [-o output_directory] [-f type] [ -c, -t ] # [ -y, -w, -e, -k, -s, -d - number ] ( -p file | -i directory ) # # == Help # # -o, --output_directory (directory) : A directory to ouput the results in. # The current directory is used if this option is not passed. # # -h, --help : This help message. # # -f, --formater (html | text) : The format to output the results in. # The default is html # # -c, --cyclo : Compute the cyclomatic complexity of the input. # # -t, --token : Count the number of tokens per line of the input. # # -y, --filter_cyclo (number) : Filter the output to only include methods # whose cyclomatic complexity are greater than the passed number. # # -w, --warn_cyclo (number) : Highlight with a warning methods whose # cyclomatic complexity are greather than or equal to the passed number. # # # -e, --error_cyclo (number) : Highligh with an error methods whose # cyclomatic complexity are greather than or equal to the passed number. # # # -k, --filter_token (number) : Filter the output to only include lines # whose token count are greater than the passed number. # # # -s, --warn_token (number) : Highlight with a warning lines whose # token count are greater than or equal to the passed number. # # # -d, --error_token (number) : Highlight with an error lines whose # token count are greater than or equal to the passed number. # # # -p, --parse_file (file) : A file to use as input. # # -i, --input_directory (directory) : All ruby files found recursively # inside the directory are passed as input. # Saikruo uses the BSD license. # # Copyright (c) 2005, Ubiquitous Business Technology (http://ubit.com) # All rights reserved. # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions are # met: # # # * Redistributions of source code must retain the above copyright # notice, this list of conditions and the following disclaimer. # # * Redistributions in binary form must reproduce the above # copyright notice, this list of conditions and the following # disclaimer in the documentation and/or other materials provided # with the distribution. # # * Neither the name of Ubiquitous Business Technology nor the names # of its contributors may be used to endorse or promote products # derived from this software without specific prior written # permission. # # # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS # "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT # LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR # A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT # OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, # SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT # LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, # DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY # THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. # # == Author # Zev Blut (zb@ubit.com) require 'irb/ruby-lex' require 'yaml' # States to watch for # once in def get the token after space, because it may also # be something like + or << for operator overloading. # Counts the number of tokens in each line. class TokenCounter include RubyToken attr_reader :current_file def initialize @files = Hash.new @tokens_per_line = Hash.new(0) @current_file = "" end # Mark file to associate with the token count. def set_current_file(file) @current_file = file @tokens_per_line = Hash.new(0) @files[@current_file] = @tokens_per_line end # Iterate through all tracked files, passing the # the provided formater the token counts. def list_tokens_per_line(formater) formater.start_count(@files.size) @files.each do |fname, tok_per_line| formater.start_file(fname) tok_per_line.sort.each do |line,num| formater.line_token_count(line,num) end formater.end_file end end # Count the token for the passed line. def count_token(line_no,token) case token when TkSPACE, TkNL, TkRD_COMMENT # Do not count these as tokens when TkCOMMENT # Ignore this only for comments in a statement? # Ignore TkCOLON,TkCOLON2 and operators? like "." etc.. when TkRBRACK, TkRPAREN, TkRBRACE # Ignore the closing of an array/index/hash/paren # The opening is counted, but no more. # Thus [], () {} is counted as 1 token not 2. else # may want to filter out comments... @tokens_per_line[line_no] += 1 end end end # Main class and structure used to compute the # cyclomatic complexity of Ruby programs. class ParseState include RubyToken attr_accessor :name, :children, :complexity, :parent, :lines @@top_state = nil def ParseState.make_top_state() @@top_state = ParseState.new(nil) @@top_state.name = "__top__" @@top_state end @@token_counter = TokenCounter.new def ParseState.set_token_counter(counter) @@token_counter = counter end def ParseState.get_token_counter @@token_counter end def initialize(lexer,parent=nil) @name = "" @children = Array.new @complexity = 0 @parent = parent @lexer = lexer @run = true # To catch one line def statements, We always have one line. @lines = 0 @last_token_line_and_char = Array.new end def top_state? self == @@top_state end def lexer=(lexer) @run = true @lexer = lexer end def make_state(type,parent = nil) cstate = type.new(@lexer,self) parent.children<< cstate cstate end def calc_complexity complexity = @complexity children.each do |child| complexity += child.calc_complexity end complexity end def calc_lines lines = @lines children.each do |child| lines += child.calc_lines end lines end def compute_state(formater) if top_state? compute_state_for_global(formater) end @children.each do |s| s.compute_state(formater) end end def compute_state_for_global(formater) global_def, @children = @children.partition do |s| !s.kind_of?(ParseClass) end return if global_def.empty? gx = global_def.inject(0) { |c,s| s.calc_complexity } gl = global_def.inject(0) { |c,s| s.calc_lines } formater.start_class_compute_state("Global", "", gx, gl) global_def.each do |s| s.compute_state(formater) end formater.end_class_compute_state("") end # Count the tokens parsed if true else ignore them. def count_tokens? true end def parse while @run do tok = @lexer.token @run = false if tok.nil? if lexer_loop?(tok) STDERR.puts "Lexer loop at line : #{@lexer.line_no} char #{@lexer.char_no}." @run = false end @last_token_line_and_char<< [@lexer.line_no.to_i, @lexer.char_no.to_i, tok] if $VERBOSE puts "DEBUG: #{@lexer.line_no} #{tok.class}:#{tok.name if tok.respond_to?(:name)}" end @@token_counter.count_token(@lexer.line_no, tok) if count_tokens? parse_token(tok) end end # Ruby-Lexer can go into a loop if the file does not end with a newline. def lexer_loop?(token) return false if @last_token_line_and_char.empty? loop_flag = false last = @last_token_line_and_char.last line = last[0] char = last[1] ltok = last[2] if ( (line == @lexer.line_no.to_i) && (char == @lexer.char_no.to_i) && (ltok.class == token.class) ) # We are potentially in a loop if @last_token_line_and_char.size >= 3 loop_flag = true end else # Not in a loop so clear stack @last_token_line_and_char = Array.new end loop_flag end def do_begin_token(token) make_state(EndableParseState, self) end def do_class_token(token) make_state(ParseClass,self) end def do_module_token(token) make_state(ParseModule,self) end def do_def_token(token) make_state(ParseDef,self) end def do_constant_token(token) nil end def do_identifier_token(token) if (token.name == "__END__" && token.char_no.to_i == 0) # The Ruby code has stopped and the rest is data so cease parsing. @run = false end nil end def do_right_brace_token(token) nil end def do_end_token(token) end_debug nil end def do_block_token(token) make_state(ParseBlock,self) end def do_conditional_token(token) make_state(ParseCond,self) end def do_conditional_do_control_token(token) make_state(ParseDoCond,self) end def do_case_token(token) make_state(EndableParseState, self) end def do_one_line_conditional_token(token) # This is an if with no end @complexity += 1 #STDOUT.puts "got IF_MOD: #{self.to_yaml}" if $VERBOSE #if state.type != "class" && state.type != "def" && state.type != "cond" #STDOUT.puts "Changing IF_MOD Parent" if $VERBOSE #state = state.parent #@run = false nil end def do_else_token(token) STDOUT.puts "Ignored/Unknown Token:#{token.class}" if $VERBOSE nil end def do_comment_token(token) make_state(ParseComment, self) end def do_symbol_token(token) make_state(ParseSymbol, self) end def parse_token(token) state = nil case token when TkCLASS state = do_class_token(token) when TkMODULE state = do_module_token(token) when TkDEF state = do_def_token(token) when TkCONSTANT # Nothing to do with a constant at top level? state = do_constant_token(token) when TkIDENTIFIER,TkFID # Nothing to do at top level? state = do_identifier_token(token) when TkRBRACE # Nothing to do at top level state = do_right_brace_token(token) when TkEND state = do_end_token(token) # At top level this might be an error... when TkDO,TkfLBRACE state = do_block_token(token) when TkIF,TkUNLESS state = do_conditional_token(token) when TkWHILE,TkUNTIL,TkFOR state = do_conditional_do_control_token(token) when TkELSIF #,TkELSE @complexity += 1 when TkELSE # Else does not increase complexity when TkCASE state = do_case_token(token) when TkWHEN @complexity += 1 when TkBEGIN state = do_begin_token(token) when TkRESCUE # Maybe this should add complexity and not begin @complexity += 1 when TkIF_MOD, TkUNLESS_MOD, TkUNTIL_MOD, TkWHILE_MOD, TkQUESTION state = do_one_line_conditional_token(token) when TkNL # @lines += 1 when TkRETURN # Early returns do not increase complexity as the condition that # calls the return is the one that increases it. when TkCOMMENT state = do_comment_token(token) when TkSYMBEG state = do_symbol_token(token) when TkError STDOUT.puts "Lexer received an error for line #{@lexer.line_no} char #{@lexer.char_no}" else state = do_else_token(token) end state.parse if state end def end_debug STDOUT.puts "got an end: #{@name} in #{self.class.name}" if $VERBOSE if @parent.nil? STDOUT.puts "DEBUG: Line #{@lexer.line_no}" STDOUT.puts "DEBUG: #{@name}; #{self.class}" # to_yaml can cause an infinite loop? #STDOUT.puts "TOP: #{@@top_state.to_yaml}" #STDOUT.puts "TOP: #{@@top_state.inspect}" # This may not be an error? #exit 1 end end end # Read and consume tokens in comments until a new line. class ParseComment < ParseState # While in a comment state do not count the tokens. def count_tokens? false end def parse_token(token) if token.is_a?(TkNL) @lines += 1 @run = false end end end class ParseSymbol < ParseState def initialize(lexer, parent = nil) super STDOUT.puts "STARTING SYMBOL" if $VERBOSE end def parse_token(token) STDOUT.puts "Symbol's token is #{token.class}" if $VERBOSE # Consume the next token and stop @run = false nil end end class EndableParseState < ParseState def initialize(lexer,parent=nil) super(lexer,parent) STDOUT.puts "Starting #{self.class}" if $VERBOSE end def do_end_token(token) end_debug @run = false nil end end class ParseClass < EndableParseState def initialize(lexer,parent=nil) super(lexer,parent) @type_name = "Class" end def do_constant_token(token) @name = token.name if @name.empty? nil end def compute_state(formater) # Seperate the Module and Class Children out cnm_children, @children = @children.partition do |child| child.kind_of?(ParseClass) end formater.start_class_compute_state(@type_name,@name,self.calc_complexity,self.calc_lines) super(formater) formater.end_class_compute_state(@name) cnm_children.each do |child| child.name = @name + "::" + child.name child.compute_state(formater) end end end class ParseModule < ParseClass def initialize(lexer,parent=nil) super(lexer,parent) @type_name = "Module" end end class ParseDef < EndableParseState def initialize(lexer,parent=nil) super(lexer,parent) @complexity = 1 @looking_for_name = true @first_space = true end # This way I don't need to list all possible overload # tokens. def create_def_name(token) case token when TkSPACE # mark first space so we can stop at next space if @first_space @first_space = false else @looking_for_name = false end when TkNL,TkLPAREN,TkfLPAREN,TkSEMICOLON # we can also stop at a new line or left parenthesis @looking_for_name = false when TkDOT @name<< "." when TkCOLON2 @name<< "::" when TkASSIGN @name<< "=" when TkfLBRACK @name<< "[" when TkRBRACK @name<< "]" else begin @name<< token.name.to_s rescue Exception => err #what is this? STDOUT.puts @@token_counter.current_file STDOUT.puts @name STDOUT.puts token.inspect STDOUT.puts err.message exit 1 end end end def parse_token(token) if @looking_for_name create_def_name(token) end super(token) end def compute_state(formater) formater.def_compute_state(@name, self.calc_complexity, self.calc_lines) super(formater) end end class ParseCond < EndableParseState def initialize(lexer,parent=nil) super(lexer,parent) @complexity = 1 end end class ParseDoCond < ParseCond def initialize(lexer,parent=nil) super(lexer,parent) @looking_for_new_line = true end # Need to consume the do that can appear at the # end of these control structures. def parse_token(token) if @looking_for_new_line if token.is_a?(TkDO) nil else if token.is_a?(TkNL) @looking_for_new_line = false end super(token) end else super(token) end end end class ParseBlock < EndableParseState def initialize(lexer,parent=nil) super(lexer,parent) @complexity = 1 @lbraces = Array.new end # Because the token for a block and hash right brace is the same, # we need to track the hash left braces to determine when an end is # encountered. def parse_token(token) if token.is_a?(TkLBRACE) @lbraces.push(true) elsif token.is_a?(TkRBRACE) if @lbraces.empty? do_right_brace_token(token) #do_end_token(token) else @lbraces.pop end else super(token) end end def do_right_brace_token(token) # we are done ? what about a hash in a block :-/ @run = false nil end end # ------------ END Analyzer logic ------------------------------------ class Filter attr_accessor :limit, :error, :warn def initialize(limit = -1, error = 11, warn = 8) @limit = limit @error = error @warn = warn end def ignore?(count) count < @limit end def warn?(count) count >= @warn end def error?(count) count >= @error end end class BaseFormater attr_accessor :warnings, :errors, :current def initialize(out, filter = nil) @out = out @filter = filter reset_data end def warn_error?(num, marker) klass = "" if @filter.error?(num) klass = ' class="error"' @errors<< [@current, marker, num] elsif @filter.warn?(num) klass = ' class="warning"' @warnings<< [@current, marker, num] end klass end def reset_data @warnings = Array.new @errors = Array.new @current = "" end end class TokenCounterFormater < BaseFormater def start(new_out=nil) reset_data @out = new_out if new_out @out.puts "Token Count" end def start_count(number_of_files) @out.puts "Counting tokens for #{number_of_files} files." end def start_file(file_name) @current = file_name @out.puts "File:#{file_name}" end def line_token_count(line_number,number_of_tokens) return if @filter.ignore?(number_of_tokens) warn_error?(number_of_tokens, line_number) @out.puts "Line:#{line_number} ; Tokens : #{number_of_tokens}" end def end_file @out.puts "" end def end_count end def end end end module HTMLStyleSheet def HTMLStyleSheet.style_sheet out = StringIO.new out.puts "" out.string end def style_sheet HTMLStyleSheet.style_sheet end end class HTMLTokenCounterFormater < TokenCounterFormater include HTMLStyleSheet def start(new_out=nil) reset_data @out = new_out if new_out @out.puts "" @out.puts style_sheet @out.puts "" end def start_count(number_of_files) @out.puts "
" @out.puts "Number of files: #{number_of_files}" @out.puts "
" end def start_file(file_name) @current = file_name @out.puts "
" @out.puts "

" @out.puts "File: #{file_name}" @out.puts "

" @out.puts "" @out.puts "" end def line_token_count(line_number,number_of_tokens) return if @filter.ignore?(number_of_tokens) klass = warn_error?(number_of_tokens, line_number) @out.puts "#{number_of_tokens}" end def end_file @out.puts "
LineTokens
#{line_number}
" end def end_count end def end @out.puts "" @out.puts "" end end class ParseStateFormater < BaseFormater def start(new_out=nil) reset_data @out = new_out if new_out end def end end def start_class_compute_state(type_name,name,complexity,lines) @current = name @out.puts "-- START #{name} --" @out.puts "Type:#{type_name} Name:#{name} Complexity:#{complexity} Lines:#{lines}" end def end_class_compute_state(name) @out.puts "-- END #{name} --" end def def_compute_state(name,complexity,lines) return if @filter.ignore?(complexity) warn_error?(complexity, name) @out.puts "Type:Def Name:#{name} Complexity:#{complexity} Lines:#{lines}" end end class StateHTMLComplexityFormater < ParseStateFormater include HTMLStyleSheet def start(new_out=nil) reset_data @out = new_out if new_out @out.puts "Cyclometric Complexity" @out.puts style_sheet @out.puts "" end def end @out.puts "" @out.puts "" end def start_class_compute_state(type_name,name,complexity,lines) @current = name @out.puts "
" @out.puts "

#{type_name} : #{name}

" @out.puts "
Total Complexity: #{complexity}
" @out.puts "
Total Lines: #{lines}
" @out.puts "" @out.puts "" end def end_class_compute_state(name) @out.puts "
MethodComplexity# Lines
" @out.puts "
" end def def_compute_state(name, complexity, lines) return if @filter.ignore?(complexity) klass = warn_error?(complexity, name) @out.puts "#{name}#{complexity}#{lines}" end end module ResultIndexGenerator def summarize_errors_and_warnings(enw, header) return "" if enw.empty? f = StringIO.new erval = Hash.new { |h,k| h[k] = Array.new } wval = Hash.new { |h,k| h[k] = Array.new } enw.each do |fname, warnings, errors| errors.each do |c,m,v| erval[v] << [fname, c, m] end warnings.each do |c,m,v| wval[v] << [fname, c, m] end end f.puts "

Errors and Warnings

" f.puts "" f.puts header f.puts print_summary_table_rows(erval, "error") f.puts print_summary_table_rows(wval, "warning") f.puts "
" f.string end def print_summary_table_rows(ewvals, klass_type) f = StringIO.new ewvals.sort { |a,b| b <=> a}.each do |v, vals| vals.sort.each do |fname, c, m| f.puts "#{c}#{m}" f.puts "#{v}" end end f.string end def list_analyzed_files(files) f = StringIO.new f.puts "

Analyzed Files

" f.puts "" f.string end def write_index(files, filename, title, header) return if files.empty? File.open(filename,"w") do |f| f.puts "#{title}" f.puts "#{HTMLStyleSheet.style_sheet}\n" f.puts "

#{title}

" enw = files.find_all { |fn,w,e| (!w.empty? || !e.empty?) } f.puts summarize_errors_and_warnings(enw, header) f.puts "
" f.puts list_analyzed_files(files) f.puts "" end end def write_cyclo_index(files, output_dir) header = "ClassMethodComplexity" write_index(files, "#{output_dir}/index_cyclo.html", "Index for cyclomatic complexity", header) end def write_token_index(files, output_dir) header = "FileLine #Tokens" write_index(files, "#{output_dir}/index_token.html", "Index for tokens per line", header) end end module Saikuro #Returns the path without the file def Saikuro.seperate_file_from_path(path) res = path.split("/") if res.size == 1 "" else res[0..res.size - 2].join("/") end end def Saikuro.analyze(files, state_formater, token_count_formater, output_dir) idx_states = Array.new idx_tokens = Array.new # parse each file files.each do |file| begin STDOUT.puts "Parsing #{file}" # create top state top = ParseState.make_top_state STDOUT.puts "TOP State made" if $VERBOSE token_counter = TokenCounter.new ParseState.set_token_counter(token_counter) token_counter.set_current_file(file) STDOUT.puts "Setting up Lexer" if $VERBOSE lexer = RubyLex.new # Turn of this, because it aborts when a syntax error is found... lexer.exception_on_syntax_error = false lexer.set_input(File.new(file,"r")) top.lexer = lexer STDOUT.puts "Parsing" if $VERBOSE top.parse fdir_path = seperate_file_from_path(file) FileUtils.makedirs("#{output_dir}/#{fdir_path}") if state_formater # output results state_io = StringIO.new state_formater.start(state_io) top.compute_state(state_formater) state_formater.end fname = "#{file}_cyclo.html" puts "writing cyclomatic #{file}" if $VERBOSE File.open("#{output_dir}/#{fname}","w") do |f| f.write state_io.string end idx_states<< [ fname, state_formater.warnings.dup, state_formater.errors.dup, ] end if token_count_formater token_io = StringIO.new token_count_formater.start(token_io) token_counter.list_tokens_per_line(token_count_formater) token_count_formater.end fname = "#{file}_token.html" puts "writing token #{file}" if $VERBOSE File.open("#{output_dir}/#{fname}","w") do |f| f.write token_io.string end idx_tokens<< [ fname, token_count_formater.warnings.dup, token_count_formater.errors.dup, ] end rescue RubyLex::SyntaxError => synerr STDOUT.puts "Lexer error for file #{file} on line #{lexer.line_no}" STDOUT.puts "#{synerr.class.name} : #{synerr.message}" rescue StandardError => err STDOUT.puts "Error while parsing file : #{file}" STDOUT.puts err.class,err.message,err.backtrace.join("\n") rescue Exception => ex STDOUT.puts "Error while parsing file : #{file}" STDOUT.puts ex.class,ex.message,ex.backtrace.join("\n") end end [idx_states, idx_tokens] end end # Really ugly command line runner stuff here for now class SaikuroCMDLineRunner require 'stringio' require 'getoptlong' require 'fileutils' require 'find' # modification to RDoc.usage that allows main_program_file to be set # for RDoc.usage require 'saikuro/usage' RDoc::main_program_file = __FILE__ include ResultIndexGenerator def get_ruby_files(path) files = Array.new Find.find(path) do |f| if !FileTest.directory?(f) if f =~ /rb$/ files<< f end end end files end def run files = Array.new output_dir = "./" formater = "html" state_filter = Filter.new(5) token_filter = Filter.new(10, 25, 50) comp_state = comp_token = false begin opt = GetoptLong.new( ["-o","--output_directory", GetoptLong::REQUIRED_ARGUMENT], ["-h","--help", GetoptLong::NO_ARGUMENT], ["-f","--formater", GetoptLong::REQUIRED_ARGUMENT], ["-c","--cyclo", GetoptLong::NO_ARGUMENT], ["-t","--token", GetoptLong::NO_ARGUMENT], ["-y","--filter_cyclo", GetoptLong::REQUIRED_ARGUMENT], ["-k","--filter_token", GetoptLong::REQUIRED_ARGUMENT], ["-w","--warn_cyclo", GetoptLong::REQUIRED_ARGUMENT], ["-s","--warn_token", GetoptLong::REQUIRED_ARGUMENT], ["-e","--error_cyclo", GetoptLong::REQUIRED_ARGUMENT], ["-d","--error_token", GetoptLong::REQUIRED_ARGUMENT], ["-p","--parse_file", GetoptLong::REQUIRED_ARGUMENT], ["-i","--input_directory", GetoptLong::REQUIRED_ARGUMENT], ["-v","--verbose", GetoptLong::NO_ARGUMENT] ) opt.each do |arg,val| case arg when "-o" output_dir = val when "-h" RDoc.usage('help') when "-f" formater = val when "-c" comp_state = true when "-t" comp_token = true when "-k" token_filter.limit = val.to_i when "-s" token_filter.warn = val.to_i when "-d" token_filter.error = val.to_i when "-y" state_filter.limit = val.to_i when "-w" state_filter.warn = val.to_i when "-e" state_filter.error = val.to_i when "-p" files<< val when "-i" files.concat(get_ruby_files(val)) when "-v" STDOUT.puts "Verbose mode on" $VERBOSE = true end end RDoc.usage if !comp_state && !comp_token rescue => err RDoc.usage end if formater =~ /html/i state_formater = StateHTMLComplexityFormater.new(STDOUT,state_filter) token_count_formater = HTMLTokenCounterFormater.new(STDOUT,token_filter) else state_formater = ParseStateFormater.new(STDOUT,state_filter) token_count_formater = TokenCounterFormater.new(STDOUT,token_filter) end state_formater = nil if !comp_state token_count_formater = nil if !comp_token idx_states, idx_tokens = Saikuro.analyze(files, state_formater, token_count_formater, output_dir) write_cyclo_index(idx_states, output_dir) write_token_index(idx_tokens, output_dir) end end