pubid 2.0.0.pre.alpha.37 → 2.0.0.pre.alpha.39

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: aac2c57a1550852debb76008ce05c727b038be5f3aef996d62ec0147816605c9
-  data.tar.gz: 765e2a70e25788b16a40756ba09aa2036092090d84573530b7817ec378c83838
+  metadata.gz: e53b03c87c2a04f877d8eba11d5689984de9c0af641258c68c8c03377d762f12
+  data.tar.gz: 5bdad1686cf8cd3709ea1494eef8c78dd7c6974bae2802ccc669db939f788731
 SHA512:
-  metadata.gz: b635a5e6d5b648433fd716a99085f857ac73e1e926cbfc7b6b60e42ea5b65ec932f2f7510a836911774d9b471dc09835d7acfd771563e6af9a5fe01f731cfced
-  data.tar.gz: f7273bf6f87c5b1679fe9a90e07b5656b672f53c9fc3124103cfb03a28737d70bfe8c5da86cfa10facd882363503f69523ad3fb1916ae534a60b77200ca2d593
+  metadata.gz: 822faa7030a35d1e87901785b0d22bb0716c1b997985c6b9d616c1eb75f3e7fe3436ca61c427c4d8abd44d14fe2e173b7a8c5ab3bc26941bb68e75f308862965
+  data.tar.gz: e636567ecb22497f4cf0c0cafdfe68d06f83e593bc15483c168cbc25053a61221d6738ac16730f2bbe0e05c88ce6d91bb3166da7386ac1ca4004363fea390a5a

data/data/ieee/update_codes.yaml

@@ -155,7 +155,9 @@ ISO/IEC/IEEE CD P26513/D1, August 2015: ISO/IEC/IEEE P26513/D1, August 2015
 "ISO/IEC/IEEE DIS P16326/201x, December 2018": "ISO/IEC/IEEE DIS P16326-2018-12"
 "ISO/IEC/IEEE/P21839, 2019(E)": "ISO/IEC/IEEE P21839-2019"
 "ISO/IEC/IEEE CD2 P12207-2: 201x(E), February 2019": "ISO/IEC/IEEE CD2 P12207.2-2019-02"
-"ISO/IEC/IEEE P42010.WD4:2019(E)": "ISO/IEC/IEEE P42010/D-4-2019"
+# The glued WD iteration re-spells to the face that canonicalises in one
+# round ("P42010/D-4-2019" renders "P42010.2019/D4", a second round).
+"ISO/IEC/IEEE P42010.WD4:2019(E)": "ISO/IEC/IEEE P42010.2019/D4"
 # the grammar parses the raw `_CDV/V3, Month YYYY` spelling directly
 # (joint_development_ieee_format slash-stage branch) — no shim needed
 "ISO/IEC/IEEE/P16085_DIS, March 2020": "ISO/IEC/IEEE DIS P16085-2020-03"

data/lib/pubid/ieee/builder.rb

@@ -16,7 +16,7 @@ module Pubid
         "Jun" => "06", "Jul" => "07", "Aug" => "08", "Sep" => "09",
         "Oct" => "10", "Nov" => "11", "Dec" => "12",
       }.freeze
-      attr_accessor :original_input
+      attr_accessor :original_input, :parser_input
       attr_reader :identifier_class
 
       def initialize(identifier_class = Identifier)
@@ -1535,10 +1535,12 @@ module Pubid
         type_value = "Std" if type_value == "Standard"
         draft_status_value = extract_value(parsed[:draft_status])
 
-        # Handle case where parser captured number without "P" prefix
-        # Check original input to see if there was a P before the number
-        # The ieee_p_identifier rule consumes P without capturing it
-        if !type_value && original_input&.match?(/IEEE\s+P/)
+        # Handle case where parser captured number without "P" prefix.
+        # The parser sees the NORMALIZED spelling: a rewrite that moves
+        # the project P behind a re-spelled type word ("IEEE
+        # PC37.30.2/D03 Rev 8" -> "IEEE Draft PC37.30.2Rev8/D03") must
+        # not be re-read as a project marker from the raw input.
+        if !type_value && parser_input&.match?(/IEEE\s+P/)
           type_value = "P"
         end
 

data/lib/pubid/ieee/identifiers/base.rb

@@ -384,10 +384,12 @@ module Pubid
           result = PreParser.preprocess(normalized)
           return build_dual(result.parts) if result.dispatch == :dual_semicolon
         end
-        parsed = Pubid::Parg::Backend.parse(:ieee, Parser.normalize_input(normalized))
+        parser_input = Parser.normalize_input(normalized)
+        parsed = Pubid::Parg::Backend.parse(:ieee, parser_input)
         builder = Builder.new(Identifier)
         # Pass the original input string to builder for context
         builder.original_input = input
+        builder.parser_input = parser_input
         builder.build(parsed)
       end
 

data/lib/pubid/ieee/parser.rb

@@ -1295,6 +1295,60 @@ 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 three-way ISO-led list glues the language marker to the
+        # year ("ISO/IEC/IEEE 8802-1AC-2018(E)"): the spaced spelling is
+        # read as the language marker and dropped, the glued one falls
+        # into parenthetical content and prints. Re-spell the boundary.
+        # Scoped to the three-way lists: the two-way IEC/IEEE copublished
+        # form keeps its glued (E) as identity ("…:2012(E)").
+        cleaned = cleaned.gsub(
+          %r{\A((?:ISO/IEC/IEEE|IEEE/ISO/IEC|IEEE/IEC/ISO)\s+[^()]+?[^()\s])\(E\)\z},
+          '\1 (E)',
+        )
+        # The stage-first joint draft's dash part carries a bare comma
+        # year ("IEC/IEEE FDIS P60780-323, 2015"): the joint route takes
+        # no bare year there and the copublished route claims the string,
+        # so drop the non-identity year and let the stage-first route
+        # read the dash part (the rendered face is the dot human).
+        cleaned = cleaned.gsub(
+          %r{\A((?:IEC/IEEE|IEEE/IEC)\s+(?:FDIS|FCD|CDV|DIS|CD|WD|PWI|NP)\s+P\d+(?:-\d+)+),\s+\d{4}\z},
+          '\1',
+        )
+        # A two-digit draft designator spelled without its D
+        # ("IEEE P1052/08, September 2018"): the corpus face is the D
+        # ordinal ("…/D08").
+        cleaned = cleaned.gsub(
+          %r{\AIEEE\s+(P[\d.-]+)/(\d{2}),\s+([A-Za-z]+)\s+(\d{4})\z},
+          'IEEE \1/D\2, \3 \4',
+        )
+        # The revision word rides after the draft ordinal
+        # ("IEEE PC37.30.2/D03 Rev 8, October 2013"): move it onto the
+        # number and mark the document as the draft it spells.
+        cleaned = cleaned.gsub(
+          %r{\AIEEE\s+((?:P)?C?[\d.]+)/D(\d+)\s+Rev\s+(\d+),\s+([A-Za-z]+)\s+(\d{4})\z},
+          'IEEE Draft \1Rev\3/D\2, \4 \5',
+        )
+        # 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.37"
+  VERSION = "2.0.0.pre.alpha.39"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: pubid
 version: !ruby/object:Gem::Version
-  version: 2.0.0.pre.alpha.37
+  version: 2.0.0.pre.alpha.39
 platform: ruby
 authors:
 - Ribose Inc.