ripnews/trunk/ripnews/ripnews.rb

#!/usr/local/bin/ruby -w

# $Dwarf: ripnews.rb,v 1.102 2005/03/01 09:18:50 ward Exp $
# $Source$

#
# Copyright (c) 2002, 2003, 2004, 2005 Ward Wouts <ward@wouts.nl>
#
# Permission to use, copy, modify, and distribute this software for any
# purpose with or without fee is hereby granted, provided that the above
# copyright notice and this permission notice appear in all copies.
#
# THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
# WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
# MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
# ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
# WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
# ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
# OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
#

require 'date'
require 'ftools'
require 'time'
require 'getoptlong'
require 'news/article'
require 'news/newsrc'
require 'tempfile'
require 'thread'
require 'thwait'
require 'encode/uuencode'
require 'encode/yenc'

###########################################################################

###########################################################################

# memory profiling stuff
MEntry = Struct.new( "MEntry", :c, :mem )
class MEntry; def to_s() "#{c} : #{mem}"; end; end

GroupEntry = Struct.new( "GroupEntry", :c, :mem, :total )
class GroupEntry; def to_s() "#{mem}\t\t#{c} x#{total}"; end; end

def profile_mem(group)
end
def aprofile_mem(group)
	t = Thread.new {
		groups = {}
		ObjectSpace.each_object { |x|
			if not [Array,Hash].include? x.class
			e = nil
			begin
				e = MEntry.new( x.class, Marshal::dump(x).size )
			rescue TypeError # undumpable
				e = MEntry.new( x.class, 0 )
			end
			if groups.has_key? e.c
				groups[e.c].mem += e.mem
				groups[e.c].total += 1
			else
				groups[e.c] = GroupEntry.new( e.c, e.mem, 1 )
			end
		end
	}
		File.open( "mem_log", "a+" ) { |file|
			file << "Group #{group}\n"
			total = 0
			file << "bytes/class/count\n"
			groups.to_a.sort_by { |e| e[1].mem }.each { |e|
				file << "#{e[1]}\n"; total += e[1].mem }
			file << "TOTAL == #{total}\n\n"
		}
	}
	sleep 10
	t.join
end
###########################################################################

Debuglevel = 0
@tstart = Time.now

def save_file(dir, name, data)
	print "savename: #{name}\n" if Debuglevel > 1
	nname = name.gsub(/\//, "-")
	nname.sub!(/\s*$/, "")
	nname.sub!(/^[\s\.-]*/, "")
	print "nname: #{nname}\n" if Debuglevel > 1
	newname = nname
	count = 1
	d = Date.today
	date = "#{d.year}#{d.month}#{d.mday}"
	while FileTest.exists?("#{dir}/#{newname}")
		newname = "#{nname}-<#{date}.#{count}>"
		count += 1
	end
	print "name: #{newname}\n" if Debuglevel > 1

	case data.class.to_s
		when "String"
			begin
				if File.move(data, "#{dir}/#{newname}")
					print "	Saving as: '#{newname}'\n"
				else
					print "couldn't rename tempfile\n"
					return false
				end
			rescue Errno::ENOENT
				print "Caught Errno::ENOENT (save_file)\n"
				print "Error: #{$!}\n"
				print "What the *beep* happened?\n"
				return false
			end
		when "Array"
			if file = File.new("#{dir}/#{newname}", "w", 0644)
				print "	Saving as: '#{newname}'\n"
				data.collect{|i| file.print "#{i}"}
			else
				print "couldn't open file for writeing\n"
				return false
			end
		when "Tempfile"
			begin
				if File.move(data.path, "#{dir}/#{newname}")
					print "	Saving as: '#{newname}'\n"
				else
					print "couldn't rename tempfile\n"
					return false
				end
			rescue Errno::ENOENT
				print "Caught Errno::ENOENT (save_file)\n"
				print "Error: #{$!}\n"
				print "What the *beep* happened?\n"
				return false
			end
		else
			print "EEEEPS Can't save data of class: #{data.class.to_s}\n"
			return false
	end
	return true
end

def parse_options(options)
	begin
		opts = GetoptLong.new(
			[ "-I", "--include",		GetoptLong::REQUIRED_ARGUMENT ],
			[ "-c", "--configfile",		GetoptLong::REQUIRED_ARGUMENT ],
			[ "-L", "--longname",		GetoptLong::NO_ARGUMENT ],
			[ "-C", "--combinedname",	GetoptLong::NO_ARGUMENT ],
			[ "-M", "--multipart",		GetoptLong::NO_ARGUMENT ],
			[ "-s",				GetoptLong::NO_ARGUMENT ],
			[ "-S", "--singlepart",		GetoptLong::NO_ARGUMENT ],
			[ "-T", "--test",		GetoptLong::NO_ARGUMENT ],
			[ "-X", "--exclude",		GetoptLong::REQUIRED_ARGUMENT ]
		)
		opts.quiet=true

		opts.each do |opt, arg|
			options[opt] = arg
		end
	rescue GetoptLong::InvalidOption
		print "#{$!}\n"
		usage
	end
	return options
end

def usage
	print "\nUsage:\n\n"
	print "ripnews.rb [-I <pattern>] [-c <file>] [-L] [-C] [-M] [-S] [-T] [-X <pattern>]\n\n"
	print "-I <pattern>	specify an include pattern\n"
	print "-c <file>	specify an alternate configfile\n"
	print "-L		use subject as filename\n"
	print "-C		use combined filenames\n"
	print "-M		get multipart articles\n"
	print "-s		exit silently if already running\n"
	print "-S		get singlepart articles\n"
	print "-T		test mode, don't update newsrc file\n"
	print "-X <pattern>	specify an exclude pattern\n"
	exit
end

def parse_config(default = {})
	if ! default.has_key?('-s')
		print "Parsing config\n"
		print "#{default['-c']}\n"
	end
	if FileTest.readable?("#{default['-c']}")
		file = File.new("#{default['-c']}")
		lines = file.readlines
	else
		lines = []
	end

	i = 0
	group = ""
	grouparr = []
	config = {}

	lines.collect!{|x|
		x.gsub!(/\$\{HOME\}/, "#{ENV['HOME']}")
		if x =~ /^\s*INCLUDEFILE=(.*?)\s*$/i
			x = File.new($1).readlines
		end
		x
	}

	lines.flatten!

	lines.collect!{|x|
		x.sub!(/^\s*/, "")
		x.sub!(/\#.*$/, "")
		x.sub!(/\s*$/, "")
		x.gsub!(/\$\{HOME\}/, "#{ENV['HOME']}")
		x.chomp
	}
	while i < lines.length
		line = lines[i]
		while line.sub!(/\s*\\$/, "") != nil
			line << lines[i+1]
			i += 1
		end
		line.sub!(/\s*$/, "")
		i += 1
		if line =~ /^OPT_(.*?)=(.*)/
			line = "-#{$1}=#{$2}"
		end
		print "#{i}: #{line}\n" if Debuglevel > 1
		if line =~ /(.*?)\s*\+=\s*(.*)/
			if group == ""
				if default.has_key?($1)
					default[$1] << $2
				else
					default[$1] = $2
				end
			else
				grouparr.collect{|g|
					if config[g].has_key?($1)
						config[g][$1] << $2
					elsif default.has_key?($1)
						config[g][$1] = default[$1] + $2
					else
						config[g][$1] = $2
					end
				}
			end
		elsif line =~ /(.*?)\s*=\s*(.*)/
			if group == ""
				default[$1] = $2
			else
				grouparr.collect{|g|
					config[g][$1] = $2
				}
			end
		elsif line =~ /(.*?)\s*\{/
			group = $1
			grouparr = group.split('|')
			grouparr.collect{|g|
				config[g] = {} unless config.has_key?(g)
			}
		elsif line =~ /^\}$/
			default.each_key{|x|
				grouparr.collect{|g|
					config[g][x] = default[x] unless config[g].has_key?(x)
				}
			}
			group = ""
			grouparr = []
		elsif line =~ /^$/
			next
		else
			print "Error parsing config on line: #{i}\n"
			return false
		end
	end

	if group != ""
		print "Error parsing config: group not terminated on line #{i}\n"
		return false
	end

	if Debuglevel > 2
		config.each_key{|x|
			print "Group: #{x}\n"
			config[x].each_key{|y|
				print "Key: '#{y}' => Value: '#{config[x][y]}'\n"
			}
		}
	end
	return config
end

def check_config
	if @config.length == 0
		print "No configuration, nothing to do\n"
		exit
	end
	@config.each_key {|i|
		unless @config[i].has_key?("-I")
			print "No inclusions given for group #{i}. Won't match anything.\n"
		end
		@config[i]["DATADIR"] ="." unless @config[i].has_key?("DATADIR")
		@config[i]["PERMISSION"] = "0755" unless @config[i].has_key?("PERMISSION")
		if @config[i].has_key?("EXTENSIONS")
			@config[i]["-S"] = @config[i]["EXTENSIONS"]
			@config[i]["-M"] = @config[i]["EXTENSIONS"]
		end
		if @config[i].has_key?("DELEXT")
			@config[i]["-SD"] = @config[i]["DELEXT"]
			@config[i]["-MD"] = @config[i]["DELEXT"]
		end
		@config[i]["-M"] = "(?!.*)" if @config[i].has_key?("-S") and ! @config[i].has_key?("-M")
		@config[i]["-S"] = "(?!.*)" if @config[i].has_key?("-M") and ! @config[i].has_key?("-S")
	}
end

def lock
	group = @config.keys[0]
	if @config[group].has_key?("LOCKFILE")
		if FileTest.exists?(@config[group]["LOCKFILE"])
			lock = File.open(@config[group]["LOCKFILE"],  "r")
			pid = lock.gets
			lock.close
			if pid
				pid.chomp!
				begin
					Process.kill(0, pid.to_i)
					if ! @config[group].has_key?('-s')
						print "Already running, exiting...\n"
					end
					exit
				rescue Errno::ESRCH
					print "Stale lock found... removing...\n"
					File.unlink(@config[group]["LOCKFILE"])
				end
			else
				print "Empty lockfile found... removing...\n"
				File.unlink(@config[group]["LOCKFILE"])
			end
		end
		lock = File.new(@config[group]["LOCKFILE"], "w")
		lock.print "#{Process.pid}\n"
		lock.close
	end
end

def unlock
	group = @config.keys[0]
	File.unlink(@config[group]["LOCKFILE"])
end

def renice
	group = @config.keys[0]
	if @config[group].has_key?("NICE")
		Process.setpriority(Process::PRIO_PROCESS, 0, @config[group]["NICE"].to_i)
	end
end

def get_single(subj, group)
	print "Fetching singlepart article: #{subj}\n"
	body = @articles.get_group_body(subj)
	if UUEncode.is_uuencoded(body)
		filename = UUEncode.get_filename(body)
		print "	filename #{filename}\n"
		unless check_ext(group, filename, "s", subj)
			print "	Skipping article...\n"
			return false
		end
		print "	UUDecoding...\n"
		mode, filename, body = UUEncode.uudecode(body)
	elsif YEnc.is_yencoded(body)
		filename = YEnc.get_filename(body)
		unless check_ext(group, filename, "s", subj)
			print "	Skipping article...\n"
			return false
		end
		print "	YDecoding...\n"
		mode, filename, body = YEnc.ydecode(body)
	else
		print "	Unknown encoding (not UU, not yEnc), skipping...\n"
		return false
	end
	if mode == false
		print "	Decoding failed skipping article...\n"
		return false
	end
	output_data(subj, mode, filename, body)
	return true
end

def get_multi(subj, group)
	print "Fetching multipart article: #{subj}\n"
	body = @articles.get_group_body_first(subj)
	if UUEncode.is_uuencoded(body) or YEnc.is_yencoded(body)
		if UUEncode.is_uuencoded(body)
			filename = UUEncode.get_filename(body)
			print "	filename #{filename}\n"
			unless check_ext(group, filename, "m", subj)
				print "	Skipping article...\n"
				return false
			end
		elsif YEnc.is_yencoded(body)
			print "yencc\n"
			filename = YEnc.get_filename(body)
			print "filename #{filename}\n"
			unless check_ext(group, filename, "m", subj)
				print "	Skipping article...\n"
				return false
			end
		end

		if @config[group]["TEMPDIR"] == nil or @config[group]["TEMPDIR"] == ""
			bodyrest = @articles.get_group_body_rest(subj)
			unless bodyrest
				print "	Skipping article...\n"
				return false
			end
			body.concat(bodyrest)
		else
			file = Tempfile.new("riptmp", @config[group]["TEMPDIR"])
			body.collect{|x| file.print "#{x}\n"}

			unless @articles.get_group_body_rest(subj, file)
				print "	Skipping article...\n"
				return false
			end
			fileout = Tempfile.new("riptmp", @config[group]["TEMPDIR"])
		end

		@decode_threads << Thread.new(body, file, fileout, subj) do |tbody, tfile, tfileout, tsubj|
			puts "inside thread pre pass\n"
			Thread.pass
			puts "inside thread post pass\n"
			if UUEncode.is_uuencoded(tbody)
				print "	UUDecoding...\n"
				if tfile
					tmode, tfilename, tbody = UUEncode.uudecode(tfile, tfileout)
				else
					tmode, tfilename, tbody = UUEncode.uudecode(tbody)
				end
			elsif YEnc.is_yencoded(tbody)
				print "	YDecoding...\n"
				begin
					if tfile
						tmode, tfilename, tbody = YEnc.ydecode(tfile, tfileout)
					else
						tmode, tfilename, tbody = YEnc.ydecode(tbody)
					end
				rescue YencError
					# XXX if there is a yenc problem I want the data so I can research it
					output_data(tsubj, 0600, "YencProblem", tbody)
					# XXX return succes even though it's not true
					Thread.current.exit
				rescue PermError
					print "#{$!}\n"
					print "	Skipping article...\n"
					Thread.current.exit
				end
			end
			if tmode == false
				print "	Decoding failed skipping article...\n"
				Thread.current.exit
			end

			if tfile
				# horrible cheat to not lose the outputted file
				tbody = tfileout.path
				tbodybase = tbody.sub(/\/[^\/]*$/, "/ripnewsdecode")
				i = 1
				while FileTest.exists?("#{tbodybase}-#{i}")
					i += 1
				end
				File.move(tbody, "#{tbodybase}-#{i}")
				tbody = "#{tbodybase}-#{i}"
				tfile.close
				tfileout.close(false)
			end
			output_data(tsubj, tmode, tfilename, tbody)
		end # thread end

		puts "ouside thread\n"

		return true
	else
		print "	Unknown encoding (not UU, not yEnc), skipping...\n"
		return false
	end
end

def fill_preselector(group)
	if @config[group].has_key?("-I")
		@articles.set_preselect_pattern(Regexp.new(@config[group]["-I"]))
	end
end

def output_data(subject, mode, filename="", body="")
	group = @articles.get_groupname
	print "	mode: #{mode}\n" if Debuglevel > 0
	print "	Filename: '#{filename}'\n" if Debuglevel > 0

	# de-crap subject...
	sub = subject.sub(/\s*$/, "")	# strip trailing spaces
	sub.sub!(/^[\s\.!-#]*/, "")	# strip leading spaces, dots, exclamation points, dashes and hashes

	# decide on a filename
	if @config[group].has_key?("-L") and @config[group]["-L"]
		print "longname\n" if Debuglevel > 1
		outfile = sub[0...@maxfilelength]
	elsif @config[group].has_key?("-C") and @config[group]["-C"]
		print "combinedname\n" if Debuglevel > 1
		outfile = sub[0...@maxfilelength-filename.length-3]
		outfile = "#{outfile} [#{filename}]"
		if outfile.length > @maxfilelength
			outfile = filename[0...@maxfilelength]
		end
	else
		print "shortname\n" if Debuglevel > 1
		outfile = filename[0...@maxfilelength]
	end

	# do the actual saving
	if save_file("#{@config[group]["DATADIR"]}/#{group}", outfile, body)
		@newsrc_lock.synchronize {
			@articles.group_update_newsrc(subject)
			@articles.save_newsrc unless @config[group].has_key?("-T") and @config[group]["-T"]
		}
	end
end

def check_ext(group, filename, mode, subject)
	case mode
	when "s"
		if @config[group].has_key?("-SD") && ( filename =~ /\.(#{@config[group]["-SD"]})$/ )
			print "Marking '#{subject}' as read\n"
			@articles.group_update_newsrc(subject)
			return false
		end
		return @config[group].has_key?("-S") ? ( filename =~ /\.(#{@config[group]["-S"]})$/ ) : true
	when "m"
		if @config[group].has_key?("-MD") && ( filename =~ /\.(#{@config[group]["-MD"]})$/ )
			print "Marking '#{subject}' as read\n"
			@articles.group_update_newsrc(subject)
			return false
		end
		return @config[group].has_key?("-M") ? ( filename =~ /\.(#{@config[group]["-M"]})$/ ) : true
	else
		print "Illegal mode \"#{mode}\" in check_ext\n"
		exit
	end
end

def get_max_file_length(tempdir=".")
	if ! FileTest.directory?("#{tempdir}") || ! FileTest.writable?("#{tempdir}")
		print "Tempdir '#{tempdir}' is not a writable directory\n"
		exit
	end
	# this is quite stupid, there is no guarantee at all the generated file names
	# don't already exist
	name = "a"*500
	name = "#$$#{name}"
	begin
		file = File.new("#{tempdir}/#{name}", "w", 0644).close
		File.delete("#{tempdir}/#{name}")
	rescue Errno::ENAMETOOLONG
		name = name[0...-1]
		retry
	rescue Errno::ENOENT
		print "#{$!}\n"
		print "strange...\n"
		retry
	end
	# this is how many characters are still likely to be appended
	# is the filename already exists '-<#{date}.#{count}>' in save_file
	# this could be brought back to 5 '-<#{count}>' ...
	return name.length - 14
end

def ward_sort(a, b)
	c = a.to_s.split(/([0-9]+)/)
	d = b.to_s.split(/([0-9]+)/)

	c.collect{|x|
		y = d.shift
		r = ((x.to_s =~ /^[0-9]+$/) && (y.to_s =~ /^[0-9]+$/)) ?
			(x.to_i <=> y.to_i) :
			(x.to_s <=> y.to_s)
		if r != 0
			return r
		end
	}
	return -1 if (d != [])
	return 0
end

def startup
	$stdout.sync=true	# line buffered output
	@defaults = {'-c' => "#{ENV['HOME']}/.ripnewsrc"}
	@defaults = parse_options(@defaults)
	@config = parse_config(@defaults)
	exit if @config == false
	check_config
	lock
	renice

	trap("HUP") {
		print "Rereading config...\n"
		config = parse_config(@defaults)
		if config != false
			@config = config
			check_config
			print "Done reading config\n"
		else
			print "Keeping old config due to errors\n"
		end
	}

	@maxfilelength = get_max_file_length(@config[@config.keys[0]]["TEMPDIR"])

	print "\n$Id$\n"
	print "Starting: #{@tstart}\n"

	if Debuglevel > 2
		@config.each_key{|i|
			print "Group: #{i}\n"
			@config[i].each_key{|j|
				print "Opt: #{j} val: #{@config[i][j]}\n"
			}
		}
	end
end

def main
	profile_mem("out side of loop still")
	for group in @config.keys.sort
		@decode_threads = []
		@newsrc_lock = Mutex.new
		profile_mem("#{group} start")
#		puts "object count:"
#		puts ObjectSpace.each_object(){}
		print "\nGetting articles for #{group}\n"
		@articles = Article.new(@config[group]["NNTPSERVER"], group, @config[group]["NEWSRCNAME"])
		fill_preselector(group)
		print "initialized\n"
		@articles.get_articles(@config[group]["CACHEDIR"])

		profile_mem("#{group} articles read")

		unless FileTest.directory?("#{@config[group]["DATADIR"]}/#{group}") or
				Dir.mkdir("#{@config[group]["DATADIR"]}/#{group}", @config[group]["PERMISSION"].oct)
			print "eeeps, couldn't create dir\n"
			exit
		end
		for i in @articles.get_group_subjects.sort{|a, b| ward_sort(a, b)}
			print "#{i}\n" if Debuglevel > 2
			if @config[group].has_key?("-MR") and i =~ /#{@config[group]["-MR"]}/
				print "Marking '#{i}' as read\n"
				@articles.group_update_newsrc(i)
				next
			end
			if !(@config[group].has_key?("-X") and i =~ /#{@config[group]["-X"]}/) and
			   i =~ /#{@config[group]["-I"]}/
				print "Match: #{i}\n" if Debuglevel > 0
				if @articles.group_is_complete(i)
					begin
						if @articles.group_is_singlepart(i)
							get_single(i, group)
						elsif @articles.group_is_multipart(i)
							get_multi(i, group)
						end
					#rescue Article::TempError, Article::PermError
					rescue TempError, PermError
						print "#{$!}\n"
						print "	Skipping article...\n"
						#print "Caught #{$!.class}\n"
						#print "Error: #{$!}\n"
						next
					end
				else
					print "Not complete: #{i}\n"
				end
			end
		end

		# hier wachten op evt. threads...
		if ! @decode_threads.empty?
			@articles.disconnect
			puts "Waiting for decode threads..."
			ThreadsWait.all_waits(@decode_threads){ |t|
				p "Thread #{t} has terminated"
			}
			puts "Decode threads all done"
		end

		@articles.quit
		@articles = nil
		profile_mem("#{group} pre-GC")
		GC.start
		profile_mem("#{group} end")
	end
end

def ending
	tend = Time.now
	print "\nFinished: #{tend}\n"
	runtime = (tend - @tstart).to_i
	h=runtime/3600
	m=runtime%3600
	s=m%60
	m=m/60
	printf("Running time: %02d:%02d:%02d\n", h, m, s)
	unlock
end

startup
main
ending