pubid 2.0.0.pre.alpha.35 → 2.0.0.pre.alpha.37

data/lib/pubid/ieee/builder.rb

@@ -1137,13 +1137,18 @@ module Pubid
         # P15026.2/D=FDIS-201008") captures the stage and the verbatim
         # date token (":2020", "-201008") with no publisher slot; the
         # designator is the IEEE face, so IEEE is the sole publisher.
-        if parsed[:iso_stage] && parsed[:draft_stage_year]
+        # The no-year form ("IEEE P12207/D=CD1") rides the same D=
+        # designator when the stage carries the glued iteration digit —
+        # the token boundary only the designation route produces.
+        if parsed[:iso_stage] && (parsed[:draft_stage_year] ||
+                                  (original_input.to_s.include?("/D=") &&
+                                   extract_value(parsed[:iso_stage]).to_s.match?(/\A[A-Z]+\d\z/)))
           stage_word = extract_value(parsed[:iso_stage])
           attributes[:publishers] ||= ["IEEE"]
           attributes[:publisher] ||= "IEEE"
           attributes[:copublisher] ||= []
           attributes[:ieee_draft] =
-            "D=#{stage_word}#{extract_value(parsed[:draft_stage_year])}"
+            "D=#{stage_word}#{extract_value(parsed[:draft_stage_year]).to_s}"
           attributes[:typed_stage] = Pubid::Ieee.locate_stage(stage_word)
           attributes[:lead_party] = "IEEE"
           attributes.delete(:year)

data/lib/pubid/ieee/parser.rb

@@ -664,10 +664,12 @@ module Pubid
                        str("ISO/IEC") | str("IEEE"))
         # ISO stage codes: FDIS, FCD, CDV; DIS/CD with an optional round
         # digit (DIS2, CD1..CD4); WD/PWI/NP.
-        iso_stage = (str("FDIS") | str("FCD") | str("CDV") |
+        iso_stage = ((str("FDIS") >> digit.maybe) | (str("FCD") >> digit.maybe) |
+                     (str("CDV") >> digit.maybe) |
                      (str("DIS") >> digit.maybe) |
                      (str("CD") >> digit.maybe) |
-                     str("WD") | str("PWI") | str("NP")).as(:iso_stage)
+                     (str("WD") >> digit.maybe) | (str("PWI") >> digit.maybe) |
+                     (str("NP") >> digit.maybe)).as(:iso_stage)
         std_noise = (space >> str("Std")).maybe
         # The empty-string marker distinguishes the stage-less PUBLISHED
         # tree in the builder (absence of :iso_stage alone means IEEE format).
@@ -719,10 +721,12 @@ module Pubid
           # D (draft) = the ISO/IEC stage it drafts, with its iteration
           # (docs/IEEE-DRAFT-STAGES.md §1.2).
           ((slash >> str("D") >> str("=") >>
-           (str("FDIS") | str("FCD") | str("CDV") |
+           ((str("FDIS") >> digit.maybe) | (str("FCD") >> digit.maybe) |
+            (str("CDV") >> digit.maybe) |
             (str("DIS") >> digit.maybe) |
             (str("CD") >> digit.maybe) |
-            str("WD") | str("PWI") | str("NP")).as(:draft_iso_stage) >>
+            (str("WD") >> digit.maybe) | (str("PWI") >> digit.maybe) |
+            (str("NP") >> digit.maybe)).as(:draft_iso_stage) >>
            (dot >> (digits >> match("[A-Za-z]").repeat(0, 1)).as(:draft_iso_iteration)).maybe) |
           (slash >> str("D") >> dash.maybe >>
            match('[0-9.]').repeat(1).as(:draft_version) >>
@@ -806,10 +810,12 @@ module Pubid
           # can't match inside a longer token and a glued no-space date
           # (".DISMay2013") is left to fall through.
           dot >>
-          (str("FDIS") | str("FCD") | str("CDV") |
+          ((str("FDIS") >> digit.maybe) | (str("FCD") >> digit.maybe) |
+           (str("CDV") >> digit.maybe) |
            (str("DIS") >> digit.maybe) |
            (str("CD") >> digit.maybe) |
-           str("WD") | str("PWI") | str("NP")).as(:iso_stage) >>
+           (str("WD") >> digit.maybe) | (str("PWI") >> digit.maybe) |
+           (str("NP") >> digit.maybe)).as(:iso_stage) >>
           (digit | match("[A-Za-z]")).absent? >>
           # optional trailing date: :YYYY, -YYYY[-MM], or text "Month YYYY"
           (

data/lib/pubid/nist/caster.rb

@@ -151,6 +151,15 @@ module Pubid
             return { second_number: value }
           end
 
+          # NEW: Handle second_number with revision_simple ("8278.r1" — the
+          # dotted revision spelling): fold to the edition, matching the
+          # glued "8278r1" canonical. The former fall-through stringified
+          # the raw parser hash into the number.
+          if type == :second_number && value.is_a?(Hash) && value[:revision_simple]
+            edition_id = value[:revision_simple][:edition_id].to_s
+            return { edition: Components::Edition.new(type: "r", id: edition_id) }
+          end
+
           # NEW: Handle second_number with revision_letter (hash with :revision_letter containing :number_only and :letter)
           # This handles "27ra" pattern where parser returns {revision_letter: {number_only: "27", letter: "a"}}
           # Should be combined to "27rA" format

data/lib/pubid/nist/parser.rb

@@ -294,8 +294,10 @@ module Pubid
              (digits >> (str(".") >> digits).maybe).as(:edition_id)).as(:revision_simple) |
             # Special patterns like "NCNR", "PERMIS", "BFRL"
             str("NCNR") | str("PERMIS") | str("BFRL") |
-            # Just capital letters (e.g., "A", "B", "C") - standalone
-            upper_letter.repeat(1, 3) |
+            # Just capital letters (e.g., "A", "B", "C") - standalone;
+            # a trailing lowercase means a capitalized word ("Suppl"),
+            # not a second number.
+            upper_letter.repeat(1, 3) >> lower_letter.absent? |
             # Regular number with optional suffix - but NOT if part of FIPS date (digit-dash-month-digit-slash)
             (digits_with_suffix >> (dash >> month_abbrev >> digits >> slash).absent?) |
             # Single lowercase letter (e.g., "a", "b") - but NOT "r" followed by digits (edition marker)
@@ -576,11 +578,20 @@ module Pubid
           (space | dash).maybe >> (digits | str("")).as(:addendum_number)).as(:addendum)
       end
 
+      # The capitalized supplement words a letter slot must yield to.
+      rule(:supplement_word) do
+        str("Suppl") | str("Supp") | str("Sup")
+      end
+
       # Supplement - enhanced to support date patterns, year patterns, and combined with revision
       # Examples: suppJan1924, supp3/1926, supp1925, supJun1925-Jun1927 (date ranges), supprev
       rule(:supplement) do
         space.maybe >>
-          (str("supp") | str("sup")) >>
+          dot.maybe >>
+          # The capitalized catalogue spellings ("NIST.SP.955.Suppl") ride
+          # the same slot; longest-first so "Suppl" is not "Sup"+"p".
+          (str("Suppl") | str("Supp") | str("Sup") |
+           str("supp") | str("sup")) >>
           (
             # Supplement followed by revision: supprev
             str("rev").as(:supplement_with_rev) |
@@ -684,14 +695,15 @@ module Pubid
           simple_series >> dot.maybe >> report_number.maybe >>
           # Edition with underscore separator (MR format: 1648_2009)
           (str("_") >> digits.as(:edition_year)).maybe >>
-          # Support letter suffix before update (e.g., 8286C-upd1) - Session 219
-          upper_letter.maybe >>
+          # Support letter suffix before update (e.g., 8286C-upd1) - Session 219;
+          # a capitalized supplement word ("Suppl") is not a letter suffix.
+          supplement_word.absent?.maybe >> upper_letter.maybe >>
           # Support revision component (r1, r5, etc.) before update
           edition.maybe >>
           update.maybe >>
           # Additional dot-separated parts (parts, version, volume, etc.)
           # MUST come before translation to avoid conflicting with language codes
-          (dot >> (digits | upper_letter)).repeat(0, 3) >>
+          (dot >> (digits | (supplement_word.absent? >> upper_letter))).repeat(0, 3) >>
           # Language codes at end (.spa, .por, .ind)
           parts.repeat >> draft.maybe
       end

data/lib/pubid/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Pubid
-  VERSION = "2.0.0.pre.alpha.35"
+  VERSION = "2.0.0.pre.alpha.37"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: pubid
 version: !ruby/object:Gem::Version
-  version: 2.0.0.pre.alpha.35
+  version: 2.0.0.pre.alpha.37
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: exe
 cert_chain: []
-date: 2026-10-19 00:00:00.000000000 Z
+date: 2026-10-20 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: lutaml-model