pubid 2.0.0.pre.alpha.36 → 2.0.0.pre.alpha.38

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"
           (
@@ -1289,6 +1295,27 @@ module Pubid
         # them to the generic standard face with a slashed copublisher.
         cleaned = cleaned.gsub(%r{((?:IEEE|ISO|IEC)(?:/(?:IEEE|ISO|IEC|CSA))+)\s+Draft\s+(P[A-Z0-9][A-Z0-9.-]*/D=[A-Z])}, '\1 \2')
         cleaned = cleaned.gsub(%r{(IEEE|ISO|IEC)//}, '\1/')
+        # The IEEE-led joint list's dash/underscore stage tail carries a
+        # month-name date ("IEEE/ISO/IEC P29119-2-DIS, December 2011",
+        # "IEEE/ISO/IEC P29119-4_FDIS, December 2020"). The generic route
+        # claims it first and answers with a slashed copublisher
+        # ("IEEE//ISO/IEC Draft …"), so re-spell the stage as the D=
+        # designator the joint route reads. Scoped to the IEEE-led list:
+        # the ISO/IEC- and IEC/IEEE-led forms already render the face.
+        cleaned = cleaned.gsub(
+          %r{\A(IEEE(?:/(?:IEEE|ISO|IEC)){2,})\s+(P[A-Z0-9][A-Z0-9.-]*?)[-_](FDIS|FCD|CDV|DIS|CD|WD|PWI|NP)(\d*),\s+([A-Za-z]+)\s+(\d{4})\z}
+        ) do
+          month = ::Pubid::Ieee::Builder::REV_MONTH_NUMBERS[
+            Regexp.last_match(5).capitalize
+          ]
+          if month
+            "#{Regexp.last_match(1)} #{Regexp.last_match(2)}/" \
+              "D=#{Regexp.last_match(3)}#{Regexp.last_match(4)}-" \
+              "#{Regexp.last_match(6)}#{month}"
+          else
+            Regexp.last_match(0)
+          end
+        end
         # The lettered REV spellings glue to the number ("P1310Rev3",
         # "P1722Rev1") or carry an after-word separator ("REV.ma",
         # "REV-ma"); the rev rules read a space or a bare word

data/lib/pubid/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Pubid
-  VERSION = "2.0.0.pre.alpha.36"
+  VERSION = "2.0.0.pre.alpha.38"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: pubid
 version: !ruby/object:Gem::Version
-  version: 2.0.0.pre.alpha.36
+  version: 2.0.0.pre.alpha.38
 platform: ruby
 authors:
 - Ribose Inc.