class Zeitwerk::Loader
Attributes
: call(String) -> void | debug(String) -> void | nil
We keep track of autoloaded directories to remove them from the registry at the end of eager loading. Files are removed as they are autoloaded, but directories need to wait due to concurrency (see why in Zeitwerk::Loader::Callbacks#on_dir_autoloaded).
: Array
Maps absolute paths for which an autoload has been set ---and not executed--- to their corresponding Zeitwerk::Cref object. "/Users/fxn/blog/app/models/user.rb" => #<Zeitwerk::Cref:... @mod=Object, @cname=:User, ...>, "/Users/fxn/blog/app/models/hotel/pricing.rb" => #<Zeitwerk::Cref:... @mod=Hotel, @cname=:Pricing, ...>, ...
: Hash[String, Zeitwerk::Cref]
: Monitor
When the path passed to Module#autoload is in the stack of features being
loaded at the moment, Ruby passes. For example, Module#autoload? returns
`nil` even if the autoload has not been attempted. See
https://bugs.ruby-lang.org/issues/21035
We call these "inceptions".
A common case is the entry point of gems managed by Zeitwerk. Their main
file is normally required and, while doing so, the loader sets an autoload
on the gem namespace. That autoload hits this edge case.
There is some logic that needs to know if an autoload for a given constant
already exists. We check Module#autoload? first, and fallback to the
inceptions just in case.
This map keeps track of pairs (cref, autoload_path) found by the loader.
The object Zeitwerk::Registry.inceptions, on the other hand, acts as a
global registry for them.
: Mutex
Maps namespace crefs to the directories that conform the namespace. When these crefs get defined we know their children are spread over those directories. We'll visit them to set up the corresponding autoloads.
A shadowed file is a file managed by this loader that is ignored when setting autoloads because its matching constant is already taken. This private set is populated lazily, as we descend. For example, if the loader has only scanned the top-level, `shadowed_files` does not have the shadowed files that may exist deep in the project tree.
: Set
If reloading is enabled, this collection maps autoload paths to their autoloaded crefs. On unload, the autoload paths are passed to callbacks, files deleted from $LOADED_FEATURES, and the crefs are deleted.
: Hash[String, Zeitwerk::Cref]
Public Class Methods
Returns an array with the absolute paths of the root directories of all registered loaders. This is a read-only collection.
: () -> Array
# File lib/zeitwerk/loader.rb, line 467 def all_dirs dirs = [] Registry.loaders.each do |loader| dirs.concat(loader.dirs) end dirs.freeze end
Broadcasts `eager_load` to all loaders. Those that have not been setup are skipped.
: () -> void
# File lib/zeitwerk/loader.rb, line 439 def eager_load_all Registry.loaders.each do |loader| begin loader.eager_load rescue SetupRequired # This is fine, we eager load what can be eager loaded. end end end
Broadcasts `eager_load_namespace` to all loaders. Those that have not been setup are skipped.
: (Module) -> void
# File lib/zeitwerk/loader.rb, line 453 def eager_load_namespace(mod) Registry.loaders.each do |loader| begin loader.eager_load_namespace(mod) rescue SetupRequired # This is fine, we eager load what can be eager loaded. end end end
This is a shortcut for require "zeitwerk" loader = Zeitwerk::Loader.new loader.tag = File.basename(__FILE__, ".rb") loader.inflector = Zeitwerk::GemInflector.new(__FILE__) loader.push_dir(__dir__) except that this method returns the same object in subsequent calls from the same file, in the unlikely case the gem wants to be able to reload. This method returns a subclass of Zeitwerk::Loader, but the exact type is private, client code can only rely on the interface.
: (?warn_on_extra_files: boolish) -> Zeitwerk::GemLoader
# File lib/zeitwerk/loader.rb, line 401 def for_gem(warn_on_extra_files: true) called_from = caller_locations(1, 1).first.path Registry.loader_for_gem(called_from, namespace: Object, warn_on_extra_files: warn_on_extra_files) end
This is a shortcut for require "zeitwerk" loader = Zeitwerk::Loader.new loader.tag = namespace.name + "-" + File.basename(__FILE__, ".rb") loader.inflector = Zeitwerk::GemInflector.new(__FILE__) loader.push_dir(__dir__, namespace: namespace) except that this method returns the same object in subsequent calls from the same file, in the unlikely case the gem wants to be able to reload. This method returns a subclass of Zeitwerk::Loader, but the exact type is private, client code can only rely on the interface.
: (Module) -> Zeitwerk::GemLoader
# File lib/zeitwerk/loader.rb, line 422 def for_gem_extension(namespace) unless namespace.is_a?(Module) # Note that Class < Module. raise Zeitwerk::Error, "#{namespace.inspect} is not a class or module object, should be" end unless real_mod_name(namespace) raise Zeitwerk::Error, "extending anonymous namespaces is unsupported" end called_from = caller_locations(1, 1).first.path Registry.loader_for_gem(called_from, namespace: namespace, warn_on_extra_files: false) end
Zeitwerk::Loader::Config::new
# File lib/zeitwerk/loader.rb, line 105 def initialize super @autoloads = {} @inceptions = Zeitwerk::Cref::Map.new @autoloaded_dirs = [] @to_unload = {} @namespace_dirs = Zeitwerk::Cref::Map.new @shadowed_files = Set.new @setup = false @eager_loaded = false @fs = FileSystem.new(self) @mutex = Mutex.new @dirs_autoload_monitor = Monitor.new Registry.loaders.register(self) end
Public Instance Methods
Returns a hash that maps the absolute paths of the managed files and directories to their respective expected constant paths.
: () -> Hash[String, String]
# File lib/zeitwerk/loader.rb, line 246 def all_expected_cpaths result = {} actual_roots.each do |root_dir, root_namespace| queue = [[root_dir, real_mod_name(root_namespace)]] while (dir, cpath = queue.shift) result[dir] = cpath prefix = cpath == "Object" ? "" : cpath + "::" @fs.ls(dir) do |basename, abspath, ftype| if ftype == :file basename.delete_suffix!(".rb") result[abspath] = "#{prefix}#{cname_for(basename, abspath)}" else if collapse?(abspath) queue << [abspath, cpath] else queue << [abspath, "#{prefix}#{cname_for(basename, abspath)}"] end end end end end result end
: (String | Pathname) -> String?
# File lib/zeitwerk/loader.rb, line 276 def cpath_expected_at(path) abspath = File.expand_path(path) raise Zeitwerk::Error.new("#{abspath} does not exist") unless File.exist?(abspath) ftype = @fs.supported_ftype?(abspath) return unless ftype return if ignored_path?(abspath) paths = [] if :file == ftype basename = File.basename(abspath, ".rb") return if @fs.hidden?(basename) paths << [basename, abspath] walk_up_from = File.dirname(abspath) else walk_up_from = abspath end root_namespace = nil @fs.walk_up(walk_up_from) do |dir| break if root_namespace = roots[dir] return if ignored_path?(dir) basename = File.basename(dir) return if @fs.hidden?(basename) paths << [basename, dir] unless collapse?(dir) end return unless root_namespace if paths.empty? real_mod_name(root_namespace) else cnames = paths.reverse_each.map { cname_for(_1, _2) } if root_namespace == Object cnames.join("::") else "#{real_mod_name(root_namespace)}::#{cnames.join("::")}" end end end
# File lib/zeitwerk/loader.rb, line 368 def log return unless logger message = yield method_name = logger.respond_to?(:debug) ? :debug : :call logger.send(method_name, "Zeitwerk@#{tag}: #{message}") end
Unloads all loaded code, and calls setup again so that the loader is able to pick any changes in the file system. This method is not thread-safe, please see how this can be achieved by client code in the README of the project.
: () -> void ! Zeitwerk::Error
# File lib/zeitwerk/loader.rb, line 232 def reload raise ReloadingDisabledError unless reloading_enabled? raise SetupRequired unless @setup unload recompute_ignored_paths recompute_collapse_dirs setup end
Sets autoloads in the root namespaces.
: () -> void
# File lib/zeitwerk/loader.rb, line 127 def setup mutex.synchronize do break if @setup actual_roots.each do |root_dir, root_namespace| define_autoloads_for_dir(root_dir, root_namespace) end on_setup_callbacks.each(&:call) @setup = true end end
# File lib/zeitwerk/loader.rb, line 363 def shadowed_file?(file) shadowed_files.member?(file) end
Removes loaded constants and configured autoloads. The objects the constants stored are no longer reachable through them. In addition, since said objects are normally not referenced from anywhere else, they are eligible for garbage collection, which would effectively unload them. This method is public but undocumented. Main interface is `reload`, which means `unload` + `setup`. This one is available to be used together with `unregister`, which is undocumented too.
: () -> void
# File lib/zeitwerk/loader.rb, line 153 def unload mutex.synchronize do raise SetupRequired unless @setup # We are going to keep track of the files that were required by our # autoloads to later remove them from $LOADED_FEATURES, thus making them # loadable by Kernel#require again. # # Directories are not stored in $LOADED_FEATURES, keeping track of files # is enough. unloaded_files = Set.new autoloads.each do |abspath, cref| if cref.autoload? unload_autoload(cref) else # Could happen if loaded with require_relative. That is unsupported, # and the constant path would escape unloadable_cpath? This is just # defensive code to clean things up as much as we are able to. unload_cref(cref) unloaded_files.add(abspath) if @fs.rb_extension?(abspath) end end to_unload.each do |abspath, cref| unless on_unload_callbacks.empty? begin value = cref.get rescue ::NameError # Perhaps the user deleted the constant by hand, or perhaps an # autoload failed to define the expected constant but the user # rescued the exception. else run_on_unload_callbacks(cref, value, abspath) end end unload_cref(cref) unloaded_files.add(abspath) if @fs.rb_extension?(abspath) end unless unloaded_files.empty? # Bootsnap decorates Kernel#require to speed it up using a cache and # this optimization does not check if $LOADED_FEATURES has the file. # # To make it aware of changes, the gem defines singleton methods in # $LOADED_FEATURES: # # https://github.com/rails/bootsnap/blob/main/lib/bootsnap/load_path_cache/core_ext/loaded_features.rb # # Rails applications may depend on bootsnap, so for unloading to work # in that setting it is preferable that we restrict our API choice to # one of those methods. $LOADED_FEATURES.reject! { |file| unloaded_files.member?(file) } end autoloads.clear autoloaded_dirs.clear to_unload.clear namespace_dirs.clear shadowed_files.clear unregister_inceptions unregister_explicit_namespaces Registry.autoloads.unregister_loader(self) @setup = false @eager_loaded = false end end
Says if the given constant path would be unloaded on reload. This predicate returns `false` if reloading is disabled. This is an undocumented method that I wrote to help transition from the classic autoloader in Rails. Its usage was removed from Rails in 7.0.
: (String) -> bool
# File lib/zeitwerk/loader.rb, line 332 def unloadable_cpath?(cpath) unloadable_cpaths.include?(cpath) end
Returns an array with the constant paths that would be unloaded on reload. This predicate returns an empty array if reloading is disabled. This is an undocumented method that I wrote to help transition from the classic autoloader in Rails. Its usage was removed from Rails in 7.0.
: () -> Array
# File lib/zeitwerk/loader.rb, line 343 def unloadable_cpaths to_unload.values.map(&:path) end
This is a dangerous method. @experimental
: () -> void
# File lib/zeitwerk/loader.rb, line 351 def unregister unregister_inceptions unregister_explicit_namespaces Registry.loaders.unregister(self) Registry.autoloads.unregister_loader(self) Registry.unregister_loader(self) end
Private Instance Methods
: (Zeitwerk::Cref, String) -> void
# File lib/zeitwerk/loader.rb, line 524 def autoload_file(cref, file) if autoload_path = cref.autoload? || Registry.inceptions.registered?(cref) # First autoload for a Ruby file wins, just ignore subsequent ones. if @fs.rb_extension?(autoload_path) shadowed_files << file log { "file #{file} is ignored because #{autoload_path} has precedence" } else promote_namespace_from_implicit_to_explicit(dir: autoload_path, file: file, cref: cref) end elsif cref.defined? shadowed_files << file log { "file #{file} is ignored because #{cref} is already defined" } else define_autoload(cref, file) end end
: (Zeitwerk::Cref) -> String?
# File lib/zeitwerk/loader.rb, line 576 def autoload_path_set_by_me_for?(cref) if autoload_path = cref.autoload? autoload_path if autoloads.key?(autoload_path) else inceptions[cref] end end
: (Zeitwerk::Cref, String) -> void
# File lib/zeitwerk/loader.rb, line 495 def autoload_subdir(cref, subdir) if autoload_path = autoload_path_set_by_me_for?(cref) if @fs.rb_extension?(autoload_path) # Scanning visited a Ruby file first, and now a directory for the same # constant has been found. This means we are dealing with an explicit # namespace whose definition was seen first. # # Registering is idempotent, and we have to keep the autoload pointing # to the file. This may run again if more directories are found later # on, no big deal. register_explicit_namespace(cref) end # If the existing autoload points to a file, it has to be preserved, if # not, it is fine as it is. In either case, we do not need to override. # Just remember the subdirectory conforms this namespace. namespace_dirs.get_or_set(cref) { [] } << subdir elsif !cref.defined? # First time we find this namespace, set an autoload for it. namespace_dirs.get_or_set(cref) { [] } << subdir define_autoload(cref, subdir) else # For whatever reason the constant that corresponds to this namespace has # already been defined, we have to recurse. log { "the namespace #{cref} already exists, descending into #{subdir}" } define_autoloads_for_dir(subdir, cref.get) end end
: (Zeitwerk::Cref, String) -> void
# File lib/zeitwerk/loader.rb, line 558 def define_autoload(cref, abspath) cref.autoload(abspath) if logger if @fs.rb_extension?(abspath) log { "autoload set for #{cref}, to be loaded from #{abspath}" } else log { "autoload set for #{cref}, to be autovivified from #{abspath}" } end end autoloads[abspath] = cref Registry.autoloads.register(abspath, self) register_inception(cref, abspath) unless cref.autoload? end
: (String, Module) -> void
# File lib/zeitwerk/loader.rb, line 477 def define_autoloads_for_dir(dir, parent) @fs.ls(dir) do |basename, abspath, ftype| if ftype == :file basename.delete_suffix!(".rb") cref = Cref.new(parent, cname_for(basename, abspath)) autoload_file(cref, abspath) else if collapse?(abspath) define_autoloads_for_dir(abspath, parent) else cref = Cref.new(parent, cname_for(basename, abspath)) autoload_subdir(cref, abspath) end end end end
`dir` is the directory that would have autovivified a namespace. `file` is the file where we've found the namespace is explicitly defined.
: (dir: String, file: String, cref: Zeitwerk::Cref) -> void
# File lib/zeitwerk/loader.rb, line 545 def promote_namespace_from_implicit_to_explicit(dir:, file:, cref:) autoloads.delete(dir) Registry.autoloads.unregister(dir) log { "earlier autoload for #{cref} discarded, it is actually an explicit namespace defined in #{file}" } # Order matters: When Module#const_added is triggered by the autoload, we # don't want the namespace to be registered yet. define_autoload(cref, file) register_explicit_namespace(cref) end
: (String) -> void
# File lib/zeitwerk/loader.rb, line 609 def raise_if_conflicting_root_dir(root_dir) if loader = Registry.conflicting_root_dir?(self, root_dir) require "pp" # Needed to have pretty_inspect available. raise Error, "loader\n\n#{pretty_inspect}\n\nwants to manage directory #{root_dir}," \ " which is already managed by\n\n#{loader.pretty_inspect}\n" end end
: (Zeitwerk::Cref) -> void
# File lib/zeitwerk/loader.rb, line 585 def register_explicit_namespace(cref) Registry.explicit_namespaces.register(cref, self) end
: (Zeitwerk::Cref, String) -> void
# File lib/zeitwerk/loader.rb, line 595 def register_inception(cref, abspath) inceptions[cref] = abspath Registry.inceptions.register(cref, abspath) end
: (String, top, String) -> void
# File lib/zeitwerk/loader.rb, line 619 def run_on_unload_callbacks(cref, value, abspath) # Order matters. If present, run the most specific one. on_unload_callbacks[cref.path]&.each { |c| c.call(value, abspath) } on_unload_callbacks[:ANY]&.each { |c| c.call(cref.path, value, abspath) } end
: (Zeitwerk::Cref) -> void
# File lib/zeitwerk/loader.rb, line 626 def unload_autoload(cref) cref.remove log { "autoload for #{cref} removed" } end
: (Zeitwerk::Cref) -> void
# File lib/zeitwerk/loader.rb, line 632 def unload_cref(cref) # Let's optimistically remove_const. The way we use it, this is going to # succeed always if all is good. cref.remove rescue ::NameError # There are a few edge scenarios in which this may happen. If the constant # is gone, that is OK, anyway. else log { "#{cref} unloaded" } end
: () -> void
# File lib/zeitwerk/loader.rb, line 590 def unregister_explicit_namespaces Registry.explicit_namespaces.unregister_loader(self) end
: () -> void
# File lib/zeitwerk/loader.rb, line 601 def unregister_inceptions inceptions.each_key do |cref| Registry.inceptions.unregister(cref) end inceptions.clear end