This provider can be used as the parent class for a provider that parses and generates files. Its content must be loaded via the ‘prefetch’ method, and the file will be written when ‘flush’ is called on the provider instance. At this point, the file is written once for every provider instance.
Once the provider prefetches the data, it’s the resource’s job to copy that data over to the @is variables.
# File lib/puppet/provider/parsedfile.rb, line 394 def create @resource.class.validproperties.each do |property| if value = @resource.should(property) @property_hash[property] = value end end mark_target_modified (@resource.class.name.to_s + "_created").intern end
# File lib/puppet/provider/parsedfile.rb, line 404 def destroy # We use the method here so it marks the target as modified. self.ensure = :absent (@resource.class.name.to_s + "_deleted").intern end
# File lib/puppet/provider/parsedfile.rb, line 410 def exists? !(@property_hash[:ensure] == :absent or @property_hash[:ensure].nil?) end
Write our data to disk.
# File lib/puppet/provider/parsedfile.rb, line 415 def flush # Make sure we've got a target and name set. # If the target isn't set, then this is our first modification, so # mark it for flushing. unless @property_hash[:target] @property_hash[:target] = @resource.should(:target) || self.class.default_target self.class.modified(@property_hash[:target]) end @resource.class.key_attributes.each do |attr| @property_hash[attr] ||= @resource[attr] end self.class.flush(@property_hash) end
Retrieve the current state from disk.
# File lib/puppet/provider/parsedfile.rb, line 441 def prefetch raise Puppet::DevError, "Somehow got told to prefetch with no resource set" unless @resource self.class.prefetch(@resource[:name] => @resource) end
# File lib/puppet/provider/parsedfile.rb, line 446 def record_type @property_hash[:record_type] end
Make sure our file is backed up, but only back it up once per transaction. We cheat and rely on the fact that @records is created on each prefetch.
# File lib/puppet/provider/parsedfile.rb, line 80 def self.backup_target(target) return nil unless target_object(target).respond_to?(:backup) @backup_stats ||= {} return nil if @backup_stats[target] == @records.object_id target_object(target).backup @backup_stats[target] = @records.object_id end
# File lib/puppet/provider/parsedfile.rb, line 22 def self.clean(hash) newhash = hash.dup [:record_type, :on_disk].each do |p| newhash.delete(p) if newhash.include?(p) end newhash end
# File lib/puppet/provider/parsedfile.rb, line 31 def self.clear @target_objects.clear @records.clear end
The mode for generated files if they are newly created. No mode will be set on existing files.
@abstract Providers inheriting parsedfile can override this method
to provide a mode. The value should be suitable for File.chmod
# File lib/puppet/provider/parsedfile.rb, line 325 def self.default_mode nil end
How to handle third party headers. @api private @abstract Providers based on ParsedFile that make use of the support for
third party headers may override this method to return +true+. When this is done, headers that are matched by the native_header_regex are not written back to disk.
# File lib/puppet/provider/parsedfile.rb, line 136 def self.drop_native_header false end
# File lib/puppet/provider/parsedfile.rb, line 36 def self.filetype @filetype ||= Puppet::Util::FileType.filetype(:flat) end
# File lib/puppet/provider/parsedfile.rb, line 40 def self.filetype=(type) if type.is_a?(Class) @filetype = type elsif klass = Puppet::Util::FileType.filetype(type) @filetype = klass else raise ArgumentError, "Invalid filetype #{type}" end end
Flush all of the targets for which there are modified records. The only reason we pass a record here is so that we can add it to the stack if necessary – it’s passed from the instance calling ‘flush’.
# File lib/puppet/provider/parsedfile.rb, line 53 def self.flush(record) # Make sure this record is on the list to be flushed. unless record[:on_disk] record[:on_disk] = true @records << record # If we've just added the record, then make sure our # target will get flushed. modified(record[:target] || default_target) end return unless defined?(@modified) and ! @modified.empty? flushed = [] begin @modified.sort { |a,b| a.to_s <=> b.to_s }.uniq.each do |target| Puppet.debug "Flushing #{@resource_type.name} provider target #{target}" flushed << target flush_target(target) end ensure @modified.reject! { |t| flushed.include?(t) } end end
Flush all of the records relating to a specific target.
# File lib/puppet/provider/parsedfile.rb, line 91 def self.flush_target(target) backup_target(target) records = target_records(target).reject { |r| r[:ensure] == :absent } target_object(target).write(to_file(records)) end
Return the header placed at the top of each generated file, warning users that modifying this file manually is probably a bad idea.
# File lib/puppet/provider/parsedfile.rb, line 103 def self.header %Q{# HEADER: This file was autogenerated at #{Time.now} # HEADER: by puppet. While it can still be managed manually, it # HEADER: is definitely not recommended.\n} end
Add another type var.
# File lib/puppet/provider/parsedfile.rb, line 141 def self.initvars @records = [] @target_objects = {} @target = nil # Default to flat files @filetype ||= Puppet::Util::FileType.filetype(:flat) super end
Return a list of all of the records we can find.
# File lib/puppet/provider/parsedfile.rb, line 153 def self.instances targets.collect do |target| prefetch_target(target) end.flatten.reject { |r| skip_record?(r) }.collect do |record| new(record) end end
Match a list of catalog resources with provider instances
@api private
@param [Array<Puppet::Resource>] resources A list of resources using this class as a provider
# File lib/puppet/provider/parsedfile.rb, line 219 def self.match_providers_with_resources(resources) return unless resources matchers = resources.dup @records.each do |record| # Skip things like comments and blank lines next if skip_record?(record) if (resource = resource_for_record(record, resources)) resource.provider = new(record) elsif respond_to?(:match) if resource = match(record, matchers) matchers.delete(resource.title) record[:name] = resource[:name] resource.provider = new(record) end end end end
Override the default method with a lot more functionality.
# File lib/puppet/provider/parsedfile.rb, line 162 def self.mk_resource_methods [resource_type.validproperties, resource_type.parameters].flatten.each do |attr| attr = attr.intern define_method(attr) do # If it's not a valid field for this record type (which can happen # when different platforms support different fields), then just # return the should value, so the resource shuts up. if @property_hash[attr] or self.class.valid_attr?(self.class.name, attr) @property_hash[attr] || :absent else if defined?(@resource) @resource.should(attr) else nil end end end define_method(attr.to_s + "=") do |val| mark_target_modified @property_hash[attr] = val end end end
Mark a target as modified so we know to flush it. This only gets used within the attr= methods.
# File lib/puppet/provider/parsedfile.rb, line 195 def self.modified(target) @modified ||= [] @modified << target unless @modified.include?(target) end
An optional regular expression matched by third party headers.
For example, this can be used to filter the vixie cron headers as erronously exported by older cron versions.
@api private @abstract Providers based on ParsedFile may implement this to make it
possible to identify a header maintained by a third party tool. The provider can then allow that header to remain near the top of the written file, or remove it after composing the file content. If implemented, the function must return a Regexp object. The expression must be tailored to match exactly one third party header.
@see ::drop_native_header @note When specifying regular expressions in multiline mode, avoid
greedy repititions such as '.*' (use .*? instead). Otherwise, the provider may drop file content between sparse headers.
# File lib/puppet/provider/parsedfile.rb, line 125 def self.native_header_regex nil end
# File lib/puppet/provider/parsedfile.rb, line 431 def initialize(record) super # The 'record' could be a resource or a record, depending on how the provider # is initialized. If we got an empty property hash (probably because the resource # is just being initialized), then we want to set up some defaults. @property_hash = self.class.record?(resource[:name]) || {:record_type => self.class.name, :ensure => :absent} if @property_hash.empty? end
Retrieve all of the data from disk. There are three ways to know which files to retrieve: We might have a list of file objects already set up, there might be instances of our associated resource and they will have a path parameter set, and we will have a default path set. We need to turn those three locations into a list of files, prefetch each one, and make sure they’re associated with each appropriate resource instance.
# File lib/puppet/provider/parsedfile.rb, line 207 def self.prefetch(resources = nil) # Reset the record list. @records = prefetch_all_targets(resources) match_providers_with_resources(resources) end
# File lib/puppet/provider/parsedfile.rb, line 253 def self.prefetch_all_targets(resources) records = [] targets(resources).each do |target| records += prefetch_target(target) end records end
Prefetch an individual target.
# File lib/puppet/provider/parsedfile.rb, line 262 def self.prefetch_target(target) begin target_records = retrieve(target) rescue Puppet::Util::FileType::FileReadError => detail puts detail.backtrace if Puppet[:trace] Puppet.err "Could not prefetch #{self.resource_type.name} provider '#{self.name}' target '#{target}': #{detail}. Treating as empty" target_records = [] end target_records.each do |r| r[:on_disk] = true r[:target] = target r[:ensure] = :present end target_records = prefetch_hook(target_records) if respond_to?(:prefetch_hook) raise Puppet::DevError, "Prefetching #{target} for provider #{self.name} returned nil" unless target_records target_records end
Is there an existing record with this name?
# File lib/puppet/provider/parsedfile.rb, line 286 def self.record?(name) return nil unless @records @records.find { |r| r[:name] == name } end
Look up a resource based on a parsed file record
@api private
@param [Hash<Symbol, Object>] record @param [Array<Puppet::Resource>] resources
@return [Puppet::Resource, nil] The resource if found, else nil
# File lib/puppet/provider/parsedfile.rb, line 246 def self.resource_for_record(record, resources) name = record[:name] if name resources[name] end end
Always make the resource methods.
# File lib/puppet/provider/parsedfile.rb, line 188 def self.resource_type=(resource) super mk_resource_methods end
Retrieve the text for the file. Returns nil in the unlikely event that it doesn’t exist.
# File lib/puppet/provider/parsedfile.rb, line 293 def self.retrieve(path) # XXX We need to be doing something special here in case of failure. text = target_object(path).read if text.nil? or text == "" # there is no file return [] else # Set the target, for logging. old = @target begin @target = path return self.parse(text) rescue Puppet::Error => detail detail.file = @target if detail.respond_to?(:file=) raise detail ensure @target = old end end end
Should we skip the record? Basically, we skip text records. This is only here so subclasses can override it.
# File lib/puppet/provider/parsedfile.rb, line 316 def self.skip_record?(record) record_type(record[:record_type]).text? end
Initialize the object if necessary.
# File lib/puppet/provider/parsedfile.rb, line 330 def self.target_object(target) # only send the default mode if the actual provider defined it, # because certain filetypes (e.g. the crontab variants) do not # expect it in their initialize method if default_mode @target_objects[target] ||= filetype.new(target, default_mode) else @target_objects[target] ||= filetype.new(target) end @target_objects[target] end
Find all of the records for a given target
# File lib/puppet/provider/parsedfile.rb, line 344 def self.target_records(target) @records.find_all { |r| r[:target] == target } end
Find a list of all of the targets that we should be reading. This is used to figure out what targets we need to prefetch.
# File lib/puppet/provider/parsedfile.rb, line 350 def self.targets(resources = nil) targets = [] # First get the default target raise Puppet::DevError, "Parsed Providers must define a default target" unless self.default_target targets << self.default_target # Then get each of the file objects targets += @target_objects.keys # Lastly, check the file from any resource instances if resources resources.each do |name, resource| if value = resource.should(:target) targets << value end end end targets.uniq.compact end
Compose file contents from the set of records.
If self.native_header_regex is not nil, possible vendor headers are identified by matching the return value against the expression. If one (or several consecutive) such headers, are found, they are either moved in front of the self.header if self.drop_native_header is false (this is the default), or removed from the return value otherwise.
@api private
# File lib/puppet/provider/parsedfile.rb, line 380 def self.to_file(records) text = super if native_header_regex and (match = text.match(native_header_regex)) if drop_native_header # concatenate the text in front of and after the native header text = match.pre_match + match.post_match else native_header = match[0] return native_header + header + match.pre_match + match.post_match end end header + text end