pubid 2.0.0.pre.alpha.9 → 2.0.0.pre.alpha.12

data/lib/pubid/tgpp/urn_generator.rb

@@ -4,16 +4,35 @@ module Pubid
   module Tgpp
     # Emits `urn:3gpp:<type>:<code>:<release>:<version>`, e.g.
     # `urn:3gpp:ts:23.207:REL-4:2.0.0`. `<code>` is number+suffix+parts.
+    #
+    # A partial identifier has no release, no version, or neither. Only the
+    # TRAILING empty segments are dropped, so a bare reference gets the clean
+    # `urn:3gpp:ts:23.207` rather than a malformed `urn:3gpp:ts:23.207::`. An
+    # INTERIOR empty segment is kept, so the release-less form keeps the URN it
+    # has always emitted: `TS 29.215/2.0.0` -> `urn:3gpp:ts:29.215::2.0.0`.
     class UrnGenerator < Pubid::UrnGenerator::Base
       def generate
-        [
-          "urn",
-          "3gpp",
-          identifier.type_prefix.downcase,
-          identifier.code,
-          identifier.release,
-          identifier.version,
-        ].join(":")
+        (head_segments + tail_segments).join(":")
+      end
+
+      private
+
+      # The namespace, the document type and the code. Always present.
+      def head_segments
+        ["urn", "3gpp", identifier.type_prefix.downcase, identifier.code]
+      end
+
+      # The two optional qualifiers, with the trailing empty ones removed.
+      def tail_segments
+        segments = [identifier.release, identifier.version]
+        # NOTE: `empty?`, not `any?` — a bare `Array#any?` tests element
+        # truthiness, so `[nil, nil].any?` is false and would strip nothing.
+        segments.pop until segments.empty? || !blank?(segments.last)
+        segments.map(&:to_s)
+      end
+
+      def blank?(value)
+        value.nil? || value.empty?
       end
     end
   end

data/lib/pubid/tgpp/urn_parser.rb

@@ -16,12 +16,14 @@ module Pubid
       def parse_urn(urn)
         body = strip_namespace(urn)
         type, code, release, version = split_parts(body)
-        # The rare no-release form (e.g. TS 29.215/2.0.0) serialises with an
-        # empty release chunk ("urn:3gpp:ts:29.215::2.0.0"); rebuild it without
-        # the release colon.
-        head = "#{type.upcase} #{code}"
-        head += ":#{release}" unless release.nil? || release.empty?
-        flavor_parse("#{head}/#{version}")
+        # A partial identifier drops its trailing segments, and the no-release
+        # form (e.g. TS 29.215/2.0.0) serialises with an empty release chunk
+        # ("urn:3gpp:ts:29.215::2.0.0"). Rebuild only the segments present, so
+        # neither the release colon nor the version slash dangles.
+        result = "#{type.upcase} #{code}"
+        result += ":#{release}" unless release.nil? || release.empty?
+        result += "/#{version}" unless version.nil? || version.empty?
+        flavor_parse(result)
       end
     end
   end

data/lib/pubid/tgpp.rb

@@ -1,5 +1,6 @@
 # frozen_string_literal: true
 
+require "pubid"
 module Pubid
   # 3GPP (3rd Generation Partnership Project) flavor.
   #
@@ -25,8 +26,13 @@ module Pubid
 
     # Parse a 3GPP identifier string
     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 ArgumentError, Pubid::INPUT_TOO_LONG_MESSAGE
+        raise Pubid::Errors::InvalidInputError, Pubid::INPUT_TOO_LONG_MESSAGE
       end
 
       Identifier.parse(identifier)

data/lib/pubid/type_resolver.rb

@@ -1,5 +1,6 @@
 # frozen_string_literal: true
 
+require "pubid"
 module Pubid
   # Resolves a polymorphic `_type` string (e.g. "pubid:iso:technical-report")
   # to the concrete Ruby class that owns it.
@@ -30,7 +31,8 @@ module Pubid
         # populated their Identifier base's polymorphic_type_map.
         ::Pubid.eager_load_flavors!
 
-        flavor = ::Pubid::Registry.get(flavor_name)
+        flavor = ::Pubid::Registry.get(flavor_name) ||
+          flavor_by_module_segment(flavor_name)
         return nil unless flavor
 
         identifier_base = flavor::Identifier
@@ -43,6 +45,19 @@ module Pubid
 
       private
 
+      # The registered flavor whose module constant gives +segment+.
+      # Identifier.polymorphic_name takes the segment from the module name
+      # ("Pubid::CenCenelec" -> "cencenelec"), and a flavor can be registered
+      # under other names ("cen_cenelec", "cen"). Without this fallback no
+      # CEN/CENELEC `_type` resolved, so the root class and a nested
+      # attribute of another flavor (BSI) read a CEN hash as the abstract
+      # root.
+      def flavor_by_module_segment(segment)
+        ::Pubid::Registry.flavors.each_value.find do |flavor_module|
+          flavor_module.name.to_s.split("::")[1]&.downcase == segment
+        end
+      end
+
       # The flavor name is the segment between the leading "pubid:" and the
       # type-name, e.g. "iso" in "pubid:iso:technical-report".
       def flavor_segment(type)

data/lib/pubid/un/identifier.rb

@@ -23,14 +23,17 @@ module Pubid
       end
 
       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 ArgumentError, Pubid::INPUT_TOO_LONG_MESSAGE
+          raise Pubid::Errors::InvalidInputError, Pubid::INPUT_TOO_LONG_MESSAGE
         end
 
         parsed = Parser.parse(identifier)
         Builder.build(parsed)
-      rescue Parslet::ParseFailed => e
-        raise "Failed to parse UN identifier '#{identifier}': #{e.message}"
       end
     end
   end

data/lib/pubid/un/parser.rb

@@ -4,7 +4,7 @@ require "parslet"
 
 module Pubid
   module Un
-    class Parser < Parslet::Parser
+    class Parser < ::Pubid::Parser::Grammar
       rule(:slash) { str("/") }
       rule(:un_prefix) { (str("UN") >> str(" ")).maybe }
       rule(:token) { match("[A-Z0-9.]").repeat(1) }

data/lib/pubid/un.rb

@@ -1,5 +1,6 @@
 # frozen_string_literal: true
 
+require "pubid"
 module Pubid
   module Un
     extend Pubid::PrefixesSupport

data/lib/pubid/urn_parser/errors.rb

@@ -1,9 +1,16 @@
 # frozen_string_literal: true
 
+require_relative "../errors"
+
 module Pubid
   module UrnParser
+    # Deprecated. The URN error moved to {Pubid::Errors::UrnParseError}, so
+    # that every pubid failure lives in one namespace under the marker module
+    # {Pubid::Errors::Error}. This alias keeps the nine `raise` sites in the
+    # flavor URN parsers, and any consumer rescuing the old name, working
+    # unchanged.
     module Errors
-      class ParseError < StandardError; end
+      ParseError = ::Pubid::Errors::UrnParseError
     end
   end
 end

data/lib/pubid/urn_parser.rb

@@ -1,5 +1,6 @@
 # frozen_string_literal: true
 
+require "pubid"
 module Pubid
   # Parses URN strings back into Identifier objects.
   #

data/lib/pubid/utils.rb

@@ -1,5 +1,6 @@
 # frozen_string_literal: true
 
+require "pubid"
 module Pubid
   module Utils
     autoload :StringNormalizer, "pubid/utils/string_normalizer"

data/lib/pubid/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Pubid
-  VERSION = "2.0.0.pre.alpha.9"
+  VERSION = "2.0.0.pre.alpha.12"
 end

data/lib/pubid/w3c/builder.rb

@@ -3,7 +3,7 @@
 module Pubid
   module W3c
     # Turns the parse tree into a concrete identifier: picks the class from the
-    # leading maturity token and splits the captured remainder into code + date.
+    # leading maturity token and splits the captured remainder into slug + date.
     class Builder
       # Maturity token -> concrete identifier class name (resolved lazily so the
       # Identifiers namespace need not be loaded at class-definition time).
@@ -22,7 +22,7 @@ module Pubid
 
       # A trailing "-<digits>" group is the date only when the digit run is
       # exactly 8 (YYYYMMDD), 6 (legacy YYMMDD) or 4 (legacy MMDD) wide. No real
-      # W3C code ends in such a run, so this never mis-splits a code that merely
+      # W3C slug ends in such a run, so this never mis-splits a slug that merely
       # ends in a short digit run (e.g. "url-1", "ATAG10").
       DATE_RE = /\A(.+)-(\d{8}|\d{6}|\d{4})\z/
 
@@ -31,8 +31,8 @@ module Pubid
       end
 
       def build(data)
-        code, date = split_date(data[:rest].to_s)
-        klass_for(data[:type]&.to_s).new(code: code, date: date)
+        number, date = split_date(data[:rest].to_s)
+        klass_for(data[:type]&.to_s).new(number: number, date: date)
       end
 
       private
@@ -42,7 +42,7 @@ module Pubid
         Identifiers.const_get(name)
       end
 
-      # @return [Array(String, String|nil)] the code and the optional date
+      # @return [Array(String, String|nil)] the slug and the optional date
       def split_date(rest)
         if (m = DATE_RE.match(rest))
           [m[1], m[2]]

data/lib/pubid/w3c/identifier.rb

@@ -7,13 +7,39 @@ module Pubid
     # this class (one per W3C maturity level), so a parsed W3C id is an
     # instance of Pubid::W3c::Identifier.
     #
-    # W3C identifiers are flat: a slug `code`, an optional publication `date`
+    # W3C identifiers are flat: a slug `number`, an optional publication `date`
     # (kept verbatim as a string — width varies: 8-digit YYYYMMDD, or legacy
     # 6-digit YYMMDD / 4-digit MMDD), and a document-type carried by the
     # concrete class (its `type_prefix` returns the printed token, e.g. "WD").
     class Identifier < ::Pubid::Identifier
-      # Slug, e.g. "charmod", "css-backgrounds-3", "07-WebData".
-      attribute :code, :string
+      # Document slug, e.g. "charmod", "css-backgrounds-3", "07-WebData".
+      #
+      # This is the relaton index key: Relaton::Index sorts and bsearches every
+      # row on `id.root.number.to_s`, and an empty key degrades that binary
+      # search to a linear scan over the whole index — silently, with no error.
+      # Every W3C identifier is its own root (the flavor has no supplement or
+      # wrapper type), so this attribute is always the key and must never be
+      # empty. It was named `code` before; there is no alias.
+      #
+      # It redefines the parent ::Pubid::Identifier's
+      # `attribute :number, Components::Code` as a plain :string. CLAUDE.md
+      # records that doing so on a class the concrete types INHERIT from can
+      # resolve against the parent definition nondeterministically under
+      # multi-flavor load (the IEEE lesson). That hazard does not apply to W3C's
+      # shape, and the reason is worth keeping: lutaml's `inherited` hook
+      # deep-dups the parent's attribute table into each subclass at
+      # class-definition time (serialize/initialization.rb#initialize_attrs), so
+      # a subclass holds a SNAPSHOT, not a live view. A base-level override is
+      # therefore safe iff every subclass body opens after this body has run —
+      # and Ruby resolves the superclass constant, running this file to
+      # completion, before opening any leaf body. That holds because this class
+      # lives in ONE file, in ONE class body, and is never reopened. Split it
+      # the way IEEE splits its base and a leaf can snapshot a half-built
+      # parent, silently reverting `number` to Components::Code. The `date`
+      # override below has relied on the same property since the flavor landed.
+      # spec/pubid/w3c/root_number_spec.rb pins it, and is only meaningful under
+      # the FULL suite.
+      attribute :number, :string
       # Publication date verbatim, or nil. Overrides the base Components::Date
       # attribute with a plain string (same mechanism JIS uses for `number`):
       # W3C dates are opaque digit runs that must round-trip exactly, and are
@@ -46,7 +72,7 @@ module Pubid
 
       key_value do
         map "_type", to: :_type, polymorphic_map: W3C_TYPE_MAP
-        map "code", to: :code
+        map "number", to: :number
         map "date", to: :date
       end
 
@@ -55,7 +81,7 @@ module Pubid
       # would otherwise fail serialization type validation.
       PUBLISHER = "W3C"
 
-      # The printed document-type token (e.g. "WD"). nil for the bare-code
+      # The printed document-type token (e.g. "WD"). nil for the bare-slug
       # Standard form; concrete typed subclasses override it.
       def type_prefix
         nil
@@ -83,18 +109,21 @@ module Pubid
       # Parse a W3C identifier string into an identifier object
       # @param identifier [String] The W3C identifier string to parse
       # @return [Identifier] The appropriate identifier object
-      # @raise [RuntimeError] If parsing fails
+      # @raise [Pubid::Errors::ParseError] If parsing fails
       def self.parse(identifier)
         # Reject pathological inputs before they reach the parser
         # (CodeQL rb/polynomial-redos barrier — inline .length by design).
+        unless identifier.is_a?(String)
+          raise Pubid::Errors::InvalidInputError,
+                Pubid::INPUT_NOT_A_STRING_MESSAGE
+        end
+
         if identifier.length > Pubid::MAX_INPUT_LENGTH
-          raise ArgumentError, Pubid::INPUT_TOO_LONG_MESSAGE
+          raise Pubid::Errors::InvalidInputError, Pubid::INPUT_TOO_LONG_MESSAGE
         end
 
         parsed = Parser.parse(identifier)
         Builder.build(parsed)
-      rescue Parslet::ParseFailed => e
-        raise "Failed to parse W3C identifier '#{identifier}': #{e.message}"
       end
     end
   end

data/lib/pubid/w3c/parser.rb

@@ -14,7 +14,7 @@ module Pubid
     # greedily consume the trailing date and make a `.maybe` date rule
     # unreliable. This is safe: no real W3C code ends in a same-width (4/6/8)
     # digit run, so the Builder's fixed-width trailing-date test never collides.
-    class Parser < Parslet::Parser
+    class Parser < ::Pubid::Parser::Grammar
       rule(:prefix) { str("W3C") >> str(" ") }
 
       # Known W3C maturity tokens. Order matters for shared prefixes: CRD before

data/lib/pubid/w3c/renderer.rb

@@ -19,7 +19,7 @@ module Pubid
         id = @id
         body = +""
         body << "#{id.type_prefix}-" if id.type_prefix
-        body << id.code.to_s
+        body << id.number.to_s
         body << "-#{id.date}" if id.date
 
         with_publisher?(id) ? "#{PUBLISHER} #{body}" : body

data/lib/pubid/w3c/urn_generator.rb

@@ -2,14 +2,14 @@
 
 module Pubid
   module W3c
-    # Emits `urn:w3c:<type-down>:<code>[:<date>]`, e.g.
+    # Emits `urn:w3c:<type-down>:<slug>[:<date>]`, e.g.
     #   "W3C WD-charmod-19991129" -> "urn:w3c:wd:charmod:19991129"
     #   "W3C 2dcontext"           -> "urn:w3c:2dcontext"
     class UrnGenerator < Pubid::UrnGenerator::Base
       def generate
         parts = ["urn", "w3c"]
         parts << identifier.type_prefix.downcase if identifier.type_prefix
-        parts << identifier.code.to_s
+        parts << identifier.number.to_s
         parts << identifier.date.to_s if identifier.date
         parts.join(":")
       end

data/lib/pubid/w3c.rb

@@ -1,5 +1,6 @@
 # frozen_string_literal: true
 
+require "pubid"
 module Pubid
   module W3c
     extend Pubid::PrefixesSupport

data/lib/pubid/xsf/identifier.rb

@@ -45,17 +45,21 @@ module Pubid
       # Parse an XSF identifier string into an identifier object
       # @param identifier [String] The XSF identifier string to parse
       # @return [Identifier] The appropriate identifier object
-      # @raise [ArgumentError] If the input exceeds the maximum length
-      # @raise [RuntimeError] If parsing fails
+      # @raise [Pubid::Errors::InvalidInputError]
+      #   If the input exceeds the maximum length
+      # @raise [Pubid::Errors::ParseError] If parsing fails
       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 ArgumentError, Pubid::INPUT_TOO_LONG_MESSAGE
+          raise Pubid::Errors::InvalidInputError, Pubid::INPUT_TOO_LONG_MESSAGE
         end
 
         parsed = Parser.parse(identifier)
         Builder.build(parsed)
-      rescue Parslet::ParseFailed => e
-        raise "Failed to parse XSF identifier '#{identifier}': #{e.message}"
       end
     end
   end

data/lib/pubid/xsf/parser.rb

@@ -4,15 +4,28 @@ require "parslet"
 
 module Pubid
   module Xsf
-    # Parslet grammar for XSF identifiers. One fixed shape: "XEP NNNN".
-    class Parser < Parslet::Parser
+    # Parslet grammar for XSF identifiers. Two shapes: "XEP NNNN", and the two
+    # named documents in SPECIAL_NUMBERS.
+    class Parser < ::Pubid::Parser::Grammar
+      # The XMPP extensions repository publishes two documents whose name is
+      # not a number: the editor README and the `xep-xxxx` template. Both reach
+      # pubid as real primary docids ("XEP README", "XEP xxxx"), so the number
+      # rule admits these two literals beside the digits. Exact case, and no
+      # other word: the flavor still rejects "XEP foo".
+      SPECIAL_NUMBERS = %w[README xxxx].freeze
+
       rule(:digit) { match["0-9"] }
       rule(:digits) { digit.repeat(1) }
 
+      rule(:special_number) do
+        SPECIAL_NUMBERS.map { |name| str(name) }.reduce(:|)
+      end
+
       # "XEP 0001" — publisher token, a single space, then the number kept as
-      # a string of digits (leading zeros preserved by the builder).
+      # a string of digits (leading zeros preserved by the builder), or the
+      # name of one of the two named documents.
       rule(:identifier) do
-        str("XEP") >> str(" ") >> digits.as(:number)
+        str("XEP") >> str(" ") >> (digits | special_number).as(:number)
       end
 
       rule(:root) { identifier }

data/lib/pubid/xsf.rb

@@ -1,5 +1,6 @@
 # frozen_string_literal: true
 
+require "pubid"
 module Pubid
   # XSF (XMPP Standards Foundation) flavor. Its documents are XEPs — XMPP
   # Extension Protocols — printed as "XEP NNNN" with a 4-digit zero-padded
@@ -23,8 +24,13 @@ module Pubid
 
     # Parse an XSF identifier string
     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 ArgumentError, Pubid::INPUT_TOO_LONG_MESSAGE
+        raise Pubid::Errors::InvalidInputError, Pubid::INPUT_TOO_LONG_MESSAGE
       end
 
       Identifier.parse(identifier)