pubid 2.0.0.pre.alpha.8 → 2.0.0.pre.alpha.11

data/lib/pubid/iana/identifier.rb

@@ -0,0 +1,140 @@
+# frozen_string_literal: true
+
+module Pubid
+  module Iana
+    # Base class for every IANA identifier AND the flavor's parse/create entry
+    # point. Concrete identifiers under Pubid::Iana::Identifiers descend from
+    # this class, so a parsed IANA id is an instance of Pubid::Iana::Identifier.
+    #
+    # IANA entries are protocol registries, not numbered standards: the value is
+    # a hierarchical registry slug (`registry` before the single "/", optional
+    # `sub_registry` after it). Both are kept verbatim.
+    #
+    # The top-level slug is stored as `number` on the concrete leaf
+    # (Identifiers::Registry) so it can serve as the relaton-index key; see the
+    # long comment there for why the attribute lives on the leaf. `registry` is
+    # a derived reader over it, so it is NOT a serialized key.
+    class Identifier < ::Pubid::Identifier
+      # The sub-registry slug (e.g. "lowpan_nhc"). nil for top registries; a nil
+      # default means it is dropped from the serialized hash when absent.
+      attribute :sub_registry, :string
+
+      # Polymorphic type map for lutaml::Model key_value (de)serialization: maps
+      # the single Registry subclass's polymorphic_name to its class name so a
+      # stored hash rebuilds the correct identifier type via from_hash.
+      IANA_TYPE_MAP = {
+        "pubid:iana:registry" => "Pubid::Iana::Identifiers::Registry",
+      }.freeze
+
+      # Identifiers::Registry merges its own block mapping `number` on top of
+      # this one. `number` is intentionally absent here: mapping it on the class
+      # the leaf inherits from would re-declare the parent's Components::Code
+      # attribute (see Identifiers::Registry).
+      key_value do
+        map "_type", to: :_type, polymorphic_map: IANA_TYPE_MAP
+        map "sub_registry", to: :sub_registry
+      end
+
+      # Publisher is always "IANA". A plain constant (not a `publisher` method)
+      # so it doesn't shadow the inherited lutaml `publisher` attribute, which
+      # would otherwise fail serialization type validation.
+      PUBLISHER = "IANA"
+
+      # Render-time flag stashed by #to_s; when false, the "IANA " token is
+      # dropped so callers can recover the bare index-key slug.
+      attr_reader :with_publisher
+
+      # The top-level registry slug. Derived from `number`, which is where it is
+      # stored — a plain method is safe because `registry` is no longer a lutaml
+      # attribute, so no generated accessor competes with it (the same shape as
+      # Pubid::Ietf::Identifiers::Bcp#series).
+      def registry
+        number
+      end
+
+      # The rendered registry code, e.g. "_6lowpan-parameters/lowpan_nhc".
+      def code
+        sub_registry ? "#{registry}/#{sub_registry}" : registry
+      end
+
+      # Basic string representation. Delegates to the renderer. Always emits the
+      # "IANA " token (the authoritative printed form) unless with_publisher is
+      # false, in which case the bare slug is rendered.
+      def to_s(with_publisher: true, **opts)
+        @with_publisher = with_publisher
+        render(format: :human, **opts)
+      end
+
+      # --- Machine-readable slug -------------------------------------------
+      # The shared MrString renderer composes publisher/type/number/... from the
+      # inherited attributes, none of which IANA populates: `publisher` is a
+      # constant rather than an attribute (see PUBLISHER above) and the builder
+      # sets no typed_stage. Without these two overrides every IANA identifier
+      # rendered the empty MR string, so all of them collided on `to_mr_string`
+      # and `to_slug` (issue #142) — and `to_slug` is used as a filename.
+      #
+      # Verified over the 3405 published relaton-data-iana rows: the resulting
+      # slugs are all distinct and stay inside the documented [a-z0-9._-] MR
+      # charset. Note that IANA slugs legitimately contain "_", which is MR's
+      # supplement separator, so an IANA MR is not re-parseable by
+      # Pubid::Parsers::MrString; uniqueness and filename-safety are what these
+      # buy, not MR round-tripping.
+      def mr_publisher
+        PUBLISHER.downcase
+      end
+
+      def mr_number_with_part
+        [number, sub_registry].compact.join(".").downcase
+      end
+
+      # The guard below has to cover this surface too: without it a nil number
+      # renders the bare publisher token "iana" for every affected row, i.e. one
+      # filename shared by all of them.
+      def to_mr_string
+        require_number!
+        super
+      end
+
+      # The top-level slug moved from a `registry` key into `number` with no
+      # alias, so a row serialized before that change deserializes with a nil
+      # number and NO error: lutaml ignores unknown keys, and relaton's
+      # Index::FileIO#id_supported? skips its round-trip check for concrete
+      # subclasses (every IANA id is an Identifiers::Registry). Failing loudly
+      # at render time is the only thing that surfaces a stale index.
+      #
+      # It is called from all three identity surfaces — to_s (via Renderer),
+      # to_urn (via UrnGenerator) and to_mr_string — because guarding only one
+      # still lets every stale row collide on the other two ("urn:iana:",
+      # "iana"), which is the silent overwrite this whole change exists to
+      # prevent.
+      def require_number!
+        return unless number.nil? || number.to_s.empty?
+
+        raise ArgumentError,
+              "Cannot render #{self.class} with an empty number " \
+              "(a pre-`number` index row stored the registry slug under " \
+              "`registry`; regenerate the index)"
+      end
+
+      # from_hash is the shared polymorphic dispatch on Pubid::Identifier;
+      # IANA_TYPE_MAP remains as the key_value polymorphic_map.
+
+      # Parse an IANA identifier string into an identifier object
+      # @param identifier [String] The IANA identifier string to parse
+      # @return [Identifier] The appropriate identifier object
+      def self.parse(identifier)
+        unless identifier.is_a?(String)
+          raise Pubid::Errors::InvalidInputError,
+                Pubid::INPUT_NOT_A_STRING_MESSAGE
+        end
+
+        if identifier.length > Pubid::MAX_INPUT_LENGTH
+          raise Pubid::Errors::InvalidInputError, Pubid::INPUT_TOO_LONG_MESSAGE
+        end
+
+        parsed = Parser.parse(identifier)
+        Builder.build(parsed)
+      end
+    end
+  end
+end

data/lib/pubid/iana/identifiers/registry.rb

@@ -0,0 +1,69 @@
+# frozen_string_literal: true
+
+module Pubid
+  module Iana
+    module Identifiers
+      # The sole IANA identifier type: a protocol registry (optionally with a
+      # sub-registry). Its class name yields the polymorphic `_type` tag
+      # "pubid:iana:registry".
+      #
+      # Examples:
+      #   IANA _6lowpan-parameters
+      #   IANA _6lowpan-parameters/lowpan_nhc
+      class Registry < Identifier
+        # The top-level registry slug ("_6lowpan-parameters"), and the index
+        # key: relaton-index bsearches on `id.root.number.to_s`, and every IANA
+        # identifier is its own root (the flavor has no wrapper or supplement
+        # types), so an empty key silently defeats the search for every row.
+        #
+        # An IANA identifier carries no numeric document number — digits only
+        # ever appear glued inside a slug token (`sip-parameters-13`,
+        # `idna-tables-11.0.0`), never as a separable component — so the slug
+        # itself is the key, exactly as an Internet-Draft's slug is
+        # (lib/pubid/ietf/identifiers/serialization.rb). Keying on the TOP-LEVEL
+        # slug rather than the full `registry/sub_registry` code means a
+        # registry and all of its sub-registries cluster into one bucket, the
+        # analogue of a draft clustering with its versions.
+        #
+        # WHY THIS IS DECLARED HERE AND NOT ON Pubid::Iana::Identifier:
+        # it redefines the parent ::Pubid::Identifier's
+        # `attribute :number, Components::Code` as a plain :string. Declaring
+        # that on a class the concrete type INHERITS from resolves against the
+        # parent definition nondeterministically under multi-flavor load — it
+        # passes an IANA-only run and can flip `root.number` to "" under the
+        # full suite (the IEEE/IETF lesson; see
+        # lib/pubid/ieee/identifiers/code_number.rb and CLAUDE.md). Declaring it
+        # directly on the leaf removes the ambiguity at the point of use.
+        # For the same reason, never define a `#number` *method* — it would
+        # collide with lutaml's generated accessor.
+        # spec/pubid/iana/root_number_spec.rb guards this and must run under the
+        # FULL suite to be meaningful.
+        attribute :number, :string
+
+        # Merged by lutaml with the base's key_value block, which maps `_type`
+        # and `sub_registry`. `registry` is deliberately NOT stored — it is the
+        # same string as `number`, and the base exposes it as a derived reader.
+        key_value do
+          map "number", to: :number
+        end
+
+        # IANA registries carry no revision/stage lifecycle; a single published
+        # typed stage keeps parity with the other flavors' type registries.
+        TYPED_STAGES = [
+          Pubid::Components::TypedStage.new(
+            code: :pubiana,
+            stage_code: :published,
+            type_code: :iana,
+            abbr: ["IANA"],
+            name: "IANA Registry",
+            harmonized_stages: [],
+          ),
+        ].freeze
+
+        def self.type
+          { key: :iana, web: :registry, title: "IANA Registry", short: "IANA" }
+        end
+      end
+    end
+  end
+end

data/lib/pubid/iana/identifiers.rb

@@ -0,0 +1,9 @@
+# frozen_string_literal: true
+
+module Pubid
+  module Iana
+    module Identifiers
+      autoload :Registry, "#{__dir__}/identifiers/registry"
+    end
+  end
+end

data/lib/pubid/iana/parser.rb

@@ -0,0 +1,38 @@
+# frozen_string_literal: true
+
+require "parslet"
+
+module Pubid
+  module Iana
+    # Parslet grammar for IANA registry identifiers.
+    #
+    # Accepts the authoritative printed form "IANA <slug>[/<sub-slug>]" as well
+    # as a bare index-key slug (no "IANA " prefix). The optional prefix is
+    # consumed and discarded — the renderer always re-emits it.
+    #
+    # Slugs use only [a-zA-Z0-9._-] and contain at most one "/", so a plain
+    # character-class repeat parses them; there is no ReDoS surface.
+    class Parser < ::Pubid::Parser::Grammar
+      # A single registry slug: letters, digits, dot, underscore, hyphen.
+      rule(:slug) { match['a-zA-Z0-9._\-'].repeat(1) }
+
+      # Optional publisher prefix. The trailing space is required, so a
+      # space-free slug never triggers it; PEG `.maybe` backtracks the whole
+      # sequence if the space is absent (e.g. a slug that happens to start
+      # "IANA...").
+      rule(:iana_prefix) { (str("IANA") >> str(" ")).maybe }
+
+      rule(:identifier) do
+        iana_prefix >>
+          slug.as(:registry) >>
+          (str("/") >> slug.as(:sub_registry)).maybe
+      end
+
+      rule(:root) { identifier }
+
+      def self.parse(input)
+        new.parse(input)
+      end
+    end
+  end
+end

data/lib/pubid/iana/renderer.rb

@@ -0,0 +1,32 @@
+# frozen_string_literal: true
+
+module Pubid
+  module Iana
+    # Human-readable renderer for IANA registry identifiers.
+    #
+    # Produces the authoritative printed form:
+    #   "IANA _6lowpan-parameters"
+    #   "IANA _6lowpan-parameters/lowpan_nhc"
+    #
+    # With `with_publisher: false` on the identifier it renders the bare
+    # index-key slug instead (no "IANA " token).
+    class Renderer < ::Pubid::Renderers::Base
+      PUBLISHER = "IANA"
+
+      def render(**_opts)
+        id = @id
+        # A stale (pre-`number`) row must not render as a bare "IANA " or
+        # "IANA /lowpan_nhc"; see Identifier#require_number!.
+        id.require_number!
+        code = id.code
+        with_publisher?(id) ? "#{PUBLISHER} #{code}" : code
+      end
+
+      private
+
+      def with_publisher?(id)
+        id.with_publisher != false
+      end
+    end
+  end
+end

data/lib/pubid/iana/urn_generator.rb

@@ -0,0 +1,19 @@
+# frozen_string_literal: true
+
+module Pubid
+  module Iana
+    # Emits `urn:iana:<registry>[:<sub_registry>]`. Slugs contain no ":"
+    # (charset is [a-zA-Z0-9._-]), so the two levels split unambiguously.
+    class UrnGenerator < Pubid::UrnGenerator::Base
+      def generate
+        # A stale (pre-`number`) row would otherwise emit the bare "urn:iana:"
+        # for every affected id; see Identifier#require_number!.
+        identifier.require_number!
+        # `number` holds the top-level registry slug (Identifiers::Registry).
+        parts = ["urn", "iana", identifier.number]
+        parts << identifier.sub_registry if identifier.sub_registry
+        parts.join(":")
+      end
+    end
+  end
+end

data/lib/pubid/iana/urn_parser.rb

@@ -0,0 +1,24 @@
+# frozen_string_literal: true
+
+module Pubid
+  module Iana
+    # Parses IANA URNs back into identifiers.
+    #
+    # UrnGenerator emits `urn:iana:<registry>[:<sub_registry>]`; this rebuilds
+    # the bare slug and delegates to the flavor's text parser (which normalizes
+    # it back to the "IANA "-prefixed printed form).
+    #
+    # Examples:
+    # - urn:iana:calipso -> IANA calipso
+    # - urn:iana:_6lowpan-parameters:lowpan_nhc
+    #     -> IANA _6lowpan-parameters/lowpan_nhc
+    class UrnParser < Pubid::UrnParser::Base
+      def parse_urn(urn)
+        body = strip_namespace(urn)
+        registry, sub_registry = split_parts(body)
+        slug = sub_registry ? "#{registry}/#{sub_registry}" : registry
+        flavor_parse(slug)
+      end
+    end
+  end
+end

data/lib/pubid/iana.rb

@@ -0,0 +1,70 @@
+# frozen_string_literal: true
+
+require "pubid"
+module Pubid
+  module Iana
+    extend Pubid::PrefixesSupport
+
+    # Sole IANA publisher token (see the parser's optional `IANA ` prefix).
+    # IANA entries are protocol registries whose printed form always leads with
+    # the "IANA" token; routing relies on it since the raw slug carries no
+    # publisher.
+    PREFIXES = ["IANA"].freeze
+
+    autoload :Builder, "#{__dir__}/iana/builder"
+    autoload :Identifier, "#{__dir__}/iana/identifier"
+    autoload :Identifiers, "#{__dir__}/iana/identifiers"
+    autoload :Parser, "#{__dir__}/iana/parser"
+    autoload :Renderer, "#{__dir__}/iana/renderer"
+    autoload :UrnGenerator, "#{__dir__}/iana/urn_generator"
+    autoload :UrnParser, "#{__dir__}/iana/urn_parser"
+
+    # Parse an IANA identifier string
+    def self.parse(identifier)
+      Identifier.parse(identifier)
+    end
+
+    # Per-flavor format registry: inherits global formats, overrides :human
+    Identifier.format_registry = FormatRegistry.new(parent: Identifier.format_registry)
+    Identifier.format_registry.register(:human, renderer: Iana::Renderer)
+
+    # Auto-discover all identifier types from the Identifiers namespace
+    # @return [Array<Class>] identifier classes that define a self.type Hash
+    def self.identifier_types
+      @identifier_types ||= Identifiers.constants
+        .filter_map { |c| begin; Identifiers.const_get(c); rescue NameError; nil; end }
+        .select { |c| c.is_a?(Class) && c.singleton_methods(false).include?(:type) }
+        .select { |c| c.type.is_a?(Hash) }
+    end
+
+    # Build typed stage index from identifier types
+    # @return [Array<Pubid::Components::TypedStage>] all typed stages
+    def self.all_typed_stages
+      @all_typed_stages ||= identifier_types.flat_map do |klass|
+        if klass.const_defined?(:TYPED_STAGES)
+          klass.const_get(:TYPED_STAGES)
+        else
+          []
+        end
+      end
+    end
+
+    # Lookup: type code -> identifier class
+    # @param code [String, Symbol] the type key to find
+    # @return [Class, nil] the matching identifier class
+    def self.locate_type(code)
+      identifier_types.find { |t| t.type[:key].to_s == code.to_s }
+    end
+
+    # Lookup: abbreviation -> typed stage
+    # @param abbr [String, Symbol] the abbreviation to find
+    # @return [Pubid::Components::TypedStage, nil] the matching typed stage
+    def self.locate_stage(abbr)
+      abbr_str = abbr.to_s.upcase
+      all_typed_stages.find { |s| s.abbr.any? { |a| a.to_s.upcase == abbr_str } }
+    end
+  end
+end
+
+# Register IANA flavor with the registry
+Pubid::Registry.register(:iana, Pubid::Iana)