# frozen_string_literal: true require 'open-uri' require 'bzip2/ffi' require 'pp' # Various helper methods to update the database of packages # This module imports information about packages of a Linux distribution # that uses the DEB package format like Debian, Ubuntu or Mint. # # It starts by loading the Release file of a release to get # information about available components and architectures. # # Next it loads the Packages lists (prefers bz2, falls back to # .gz or even the uncompressed version). module DebImporter # Defines a Debian release class Release attr_reader :architectures, :components, :description, :codename, :origin, :version, :files # Load and parse a Release file of an APT repository def initialize(dist_url) # rubocop:disable Metrics/MethodLength @dist_url = dist_url release_url = "#{dist_url}/Release" Rails.logger.debug "Loading Release file from #{release_url}" URI.open(release_url) do |release_data| fields = get_fields(release_data) @architectures = fields[:Architectures] @components = fields[:Components] @description = fields[:Description] @codename = fields[:Codename] @origin = fields[:Origin] @version = fields[:Version] # TODO: check sizes and checksums end end # Get package information from translation (i18n) files. # Returns an enumerator of packages. def i18n(component, language) url = "#{@dist_url}/#{component}/i18n/Translation-#{language}" file = find_and_open_compressed_url(url) if file return get_paragraphs(file) else return [] end end # Look for the file or URL in various compressed formats # (e.g. bz2, gz) and fall back to plain text format. def find_and_open_compressed_url(base_url) Rails.logger.debug "Looking for files at URL #{base_url} with different compressions" ['.bz2', '.gz', ''].each do |suffix| begin url = "#{base_url}#{suffix}" Rails.logger.debug "Checking if file at #{url} is available" file = URI.open(url) rescue OpenURI::HTTPError => e Rails.logger.debug "Loading #{url} lead to error #{e}. skipping." next end file2 = nil # Decompress file depending on its filename suffix case suffix when '.bz2' file2 = Bzip2::FFI::Reader.read(file) when '.gz' file2 = Zlib::GzipReader.new(file) else # plain text file2 = file end Rails.logger.debug "File containing translations is: #{url}" # Return an enumerator that iterates over lines of the file return file2 rescue Errno::ENOENT Rails.logger.debug "URL #{url} could not be opened. Skipping." end Rails.logger.error "No file found at #{url} and various compression extensions." end # Try to load the Packages file for a certain component (e.g. "main") # and architecture (e.g. "amd64") def packages(component, architecture) # create path like "main/binary-amd64/Packages" packages_path = "#{@dist_url}/#{component}/binary-#{architecture}/Packages" Rails.logger.debug "Loading packages from #{packages_path}" file = find_and_open_compressed_url(packages_path) get_paragraphs(file) end end private # Gather the fields of a Debian control file section and return them as a hash def get_fields(data) fields = {} name = value = '' data.each do |line| case line when /^(\S+?): (.+)/ # "Key: Value" fields[name.to_sym] = value unless value.empty? name = $1 value = $2 when /^(\S+?):$/ # "Key:" (start of multi-line entry without value in line) fields[name.to_sym] = value unless value.empty? name = $1 value = "" when /^\s(.+)/ # " Indented multi-line value" # Add a newline for multi-line entries ("Key: Value\n Foo\n Bar") value += "\n" unless value.empty? value += $1 when /^\s+$/ # Empty line break end end # Any lines left at the end of the input? fields[name.to_sym] = value unless value.empty? # pp fields # puts '------------' fields end # Iterator that splits up the input of a debian control file # by empty lines. For example Debian "Packages" files consist # of one paragraph for each package listed in it. def get_paragraphs(data) Rails.logger.debug("get_paragraphs – data: #{data.class}") Enumerator.new do |enum| gathered_lines = [] # collects all lines belonging to a field data.each_line do |line| #Rails.logger.debug("___" + line) if line.chomp.empty? # empty line found that seperates paragraphs if gathered_lines.any? # any lines gathered so far? # Rails.logger.debug('>>>>>>>>>>>>>>>>>>>>>> Gathered lines:') # Rails.logger.debug(gathered_lines.inspect) # Rails.logger.debug('<<<<<<<<<<<<<<<<<<<<<<') enum.yield get_fields(gathered_lines) gathered_lines = [] end else gathered_lines << line end end # Any lines left after the last empty line and the end of the input? enum.yield get_fields(gathered_lines) if gathered_lines.any? end end # Represents the version of a Debian package class Version attr_reader :epoch, :upstream, :revision, :version_string def initialize(version_string) @version_string = version_string # Split into "[epoch:]upstream[-revision]" # See: https://www.debian.org/doc/debian-policy/ch-controlfields.html#version unless /^((?\d+)\:)?(?.+?)(\-(?.+))?$/ =~ version_string raise ArgumentError, "Cannot parse version string: #{version_string}" end @epoch = epoch || '0' @upstream = upstream @revision = revision if revision end def to_s @version_string end def >(other) dpkg_compare_version(@version_string, other, 'gt') end def <(other) dpkg_compare_version(@version_string, other, 'lt') end private # Determine which version string denotes the newer package # v1: version of the first package # v2: version of the second package # op: the comparison operator (default: 'gt' / greater-than) # # The algorithm is described in the Debian Policy at # https://www.debian.org/doc/debian-policy/ch-controlfields.html#s-f-Version # The reference implementation is "dpkg --compare-versions …" def dpkg_compare_version(v1, v2, op) system("dpkg --compare-versions #{v1} #{op} #{v2}") end end # /class end # module