metanorma-standoc 2.5.0 → 2.5.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 1d6a56b88b9f58482764ff6aede297e389577151d0d112a215b2215d5cfb41cf
-  data.tar.gz: cb0a1a1d73930e89169a27185bcce3d19130f1150a22fd7345da61a73dbd759f
+  metadata.gz: 20b6d6b6948c9eae42b87ef8c6c8a96fcbabdd64f44d3effd972285e5f30ec62
+  data.tar.gz: c048925049205b4fbbb2ce7883bc786b7ffd97970bae250226844fbf9801000f
 SHA512:
-  metadata.gz: 158547c563e269b769f39123cbd03f749a4891f7e4c772e9462ea40a9221ee4b58839991b2580fc40d9a24007387866d2741867ee6a8f5f39770fa3d38a57138
-  data.tar.gz: ba7b26ed9ebfae0f92dedc50750aad4e31c53b5e5b16defe3e2a2ea6583049fab738f07449846ac62aa74af06944d725389fed9d54357bf9ff9c33b3196cac2f
+  metadata.gz: 437c944d237d240b0efc821cc7b84e99dc919557f8474c07922c1dd688dbdb7838e13d523fcfe44b308bdf322c1b4a5405a7c95553ef633e12b7e99b28906714
+  data.tar.gz: f233558ba151d531f9c67c07e75e6b2c22aabc38054db61d594abb17b6445c4e31a52931129686156b53e11bec282ec256d964b1bd9b339eca64449a8109a75d

data/lib/metanorma/standoc/cleanup_inline.rb

@@ -1,5 +1,6 @@
 require "metanorma-utils"
 require "digest"
+require "addressable/uri"
 
 module Metanorma
   module Standoc
@@ -208,8 +209,8 @@ module Metanorma
 
       def link_cleanup(xmldoc)
         xmldoc.xpath("//link[@target]").each do |l|
-          l["target"] = URI.parse(l["target"]).to_s
-        rescue StandardError
+          l["target"] = Addressable::URI.parse(l["target"]).to_s
+        rescue Addressable::URI::InvalidURIError
           err = "Malformed URI: #{l['target']}"
           @log.add("Anchors", l, err)
           @fatalerror << err

data/lib/metanorma/standoc/converter.rb

@@ -12,7 +12,7 @@ require_relative "validate"
 require_relative "utils"
 require_relative "cleanup"
 require_relative "reqt"
-require_relative "./macros"
+require_relative "macros"
 
 module Metanorma
   module Standoc
@@ -22,6 +22,7 @@ module Metanorma
       Asciidoctor::Extensions.register do
         preprocessor Metanorma::Standoc::EmbedIncludeProcessor
         preprocessor Metanorma::Standoc::NamedEscapePreprocessor
+        preprocessor Metanorma::Standoc::LinkProtectPreprocessor
         preprocessor Metanorma::Standoc::Datamodel::AttributesTablePreprocessor
         preprocessor Metanorma::Standoc::Datamodel::DiagramPreprocessor
         preprocessor Metanorma::Plugin::Datastruct::Json2TextPreprocessor

data/lib/metanorma/standoc/macros.rb

@@ -1,14 +1,14 @@
 require "uuidtools"
 require "yaml"
 require "csv"
-require_relative "./macros_inline"
-require_relative "./macros_plantuml"
-require_relative "./macros_terms"
-require_relative "./macros_form"
-require_relative "./macros_note"
-require_relative "./macros_embed"
-require_relative "./datamodel/attributes_table_preprocessor"
-require_relative "./datamodel/diagram_preprocessor"
+require_relative "macros_inline"
+require_relative "macros_plantuml"
+require_relative "macros_terms"
+require_relative "macros_form"
+require_relative "macros_note"
+require_relative "macros_embed"
+require_relative "datamodel/attributes_table_preprocessor"
+require_relative "datamodel/diagram_preprocessor"
 require "metanorma-plugin-datastruct"
 require "metanorma-plugin-glossarist"
 require "metanorma-plugin-lutaml"
@@ -22,7 +22,7 @@ module Metanorma
 
       def init_indent(line)
         /^(?<prefix>[ \t]*)(?<suffix>.*)$/ =~ line
-        prefix = prefix.gsub(/\t/, "\u00a0\u00a0\u00a0\u00a0")
+        prefix = prefix.gsub("\t", "\u00a0\u00a0\u00a0\u00a0")
           .gsub(/ /, "\u00a0")
         prefix + suffix
       end
@@ -31,9 +31,9 @@ module Metanorma
         ignore = false
         lines.each_with_index do |l, i|
           /^(--+|====+|\|===|\.\.\.\.+|\*\*\*\*+|\+\+\+\++|````+|____\+)$/
-            .match(l) && (ignore = !ignore)
-          next if l.empty? || l.match(/ \+$/) || /^\[.*\]$/.match?(l) || ignore
-          next if i == lines.size - 1 ||
+            .match(l) and (ignore = !ignore)
+          next if l.empty? || l.match(/ \+$/) || /^\[.*\]$/.match?(l) ||
+            ignore || i == lines.size - 1 ||
             (i < lines.size - 1 && lines[i + 1].empty?)
 
           lines[i] += " +"
@@ -60,5 +60,111 @@ module Metanorma
         ::Asciidoctor::Reader.new lines
       end
     end
+
+    # refer https://github.com/asciidoctor/asciidoctor/blob/main/lib/asciidoctor/substitutors.rb
+    # Not using TreeProcessor because that is still too close to
+    # inline expressions being processed on access (e.g. titles)
+    class LinkProtectPreprocessor < Asciidoctor::Extensions::Preprocessor
+      def init
+        pass = true # process as passthrough: init = true until
+        # hit end of doc header
+        is_delim = false # current line is a no-substititon block delimiter
+        pass_delim = false # current line is a passthrough delimiter
+        delimln = "" # delimiter line of current block(s);
+        # init value looks for end of doc header
+        { pass: pass, is_delim: is_delim, pass_delim: pass_delim,
+          delimln: delimln }
+      end
+
+      def process(_document, reader)
+        p = init
+        lines = reader.readlines.map do |t|
+          p = pass_status(p, t.rstrip)
+          !p[:pass] && t.include?(":") and t = inlinelinkmacro(inlinelink(t))
+          t
+        end
+        ::Asciidoctor::Reader.new lines
+      end
+
+      def pass_status(status, text)
+        text == "++++" && !status[:delimln] and status[:pass] = !status[:pass]
+        if status[:is_delim] && /^(-+|\*+|=+|_+)$/.match?(text)
+          status[:delimln] = text
+          status[:pass] = true
+        elsif status[:pass_delim]
+          status[:delimln] = "" # end of paragraph for paragraph with [pass]
+        elsif status[:delimln] && text == status[:delimln]
+          status[:pass] = false
+          status[:delimln] = nil
+        end
+        status[:is_delim] = /^\[(source|listing|literal|pass)\b/.match?(text)
+        status[:pass_delim] = /^\[(pass)\b/.match?(text)
+        status
+      end
+
+      PASS_INLINE_MACROS = %w(pass pass-format identifier std-link stem)
+        .join("|").freeze
+
+      PASS_INLINE_MACRO_STR = <<~REGEX.freeze
+        (
+          \\b(?<![-\\\\])                        # word-separator, no hyphen or backslash
+          (?:                                    # don't capture these!
+            (?:#{PASS_INLINE_MACROS}):[^\\s\\[]* | # macro name, :, second key. OR:
+            span:uri \\b [^\\s\\[]*              # span:uri, third key
+          )
+          \\[.*?(?<!\\\\)\\]                     # [ ... ] not preceded by \\
+        )
+      REGEX
+      PASS_INLINE_MACRO_RX = /#{PASS_INLINE_MACRO_STR}/xo.freeze
+
+      def pass_inline_split(text)
+        text.split(PASS_INLINE_MACRO_RX).each.map do |x|
+          PASS_INLINE_MACRO_RX.match?(x) ? x : yield(x)
+        end
+      end
+
+      # InlineLinkRx = %r((^|link:|#{CG_BLANK}|&lt;|[>\(\)\[\];"'])(\\?(?:https?|file|ftp|irc)://)(?:([^\s\[\]]+)\[(|#{CC_ALL}*?[^\\])\]|([^\s\[\]<]*([^\s,.?!\[\]<\)]))))m
+      #
+      InlineLinkRx = %r((^|(?<![-\\])\blink:(?!\+)|\p{Blank}|&lt;|[<>\(\)\[\];"'])((?:https?|file|ftp|irc)://)(?:([^\s\[\]]+)(?:(\[(|.*?[^\\])\])|([^\s\[\]<]*([^\s,.?!\[\]<\)])))))m.freeze
+
+      def inlinelink(text)
+        text.include?("://") or return text
+        pass_inline_split(text) do |x|
+          inlinelink_escape(x)
+        end.join
+      end
+
+      def inlinelink_escape(text)
+        text.gsub(InlineLinkRx) do
+          body, suffix = $4.nil? ? [$3 + $6, "[]"] : [$3, ""]
+          p = $1 and s = $2 and b = $4
+          if p == "link:" then "#{p}++#{s}#{body}++#{b}#{suffix}"
+          elsif p == "<"
+            "#{p}link:++#{s}#{body.sub(/>$/, '')}++#{b}#{suffix}>"
+          else "#{p}link:++#{s}#{body}++#{b}#{suffix}"
+          end
+        end
+      end
+
+      # InlineLinkMacroRx = /\\?(?:link|(mailto)):(|[^:\s\[][^\s\[]*)\[(|#{CC_ALL}*?[^\\])\]/m
+      InlineLinkMacroRx1 = <<~REGEX.freeze
+        (\\\\?\\b(?<!-)                  # optional backslash, no hyphen, word boundary
+          (?:link|mailto):)              # link: or mailto:
+        (?!\\+)                          # no link:+ passthrough
+        (|[^:\\s\\[][^\\s\\[]*)          # link: ... up to [
+        (\\[(|.*?[^\\\\])\\])            # [ ... ], no ]
+      REGEX
+      InlineLinkMacroRx = /#{InlineLinkMacroRx1}/x.freeze
+
+      def inlinelinkmacro(text)
+        (text.include?("[") &&
+          ((text.include? "link:") || (text.include? "ilto:"))) or return text
+        pass_inline_split(text) do |x|
+          x.gsub(InlineLinkMacroRx) do
+            "#{$1}++#{$2}++#{$3}"
+          end
+        end.join
+      end
+    end
   end
 end

data/lib/metanorma/standoc/macros_inline.rb

@@ -44,7 +44,7 @@ module Metanorma
         text = attr["text"]
         text = "((#{text}))" unless /^\(\(.+\)\)$/.match?(text)
         out = parent.sub_macros(text)
-        out.sub(/<index>/, "<index to='#{target}'>")
+        out.sub("<index>", "<index to='#{target}'>")
       end
     end
 
@@ -148,7 +148,7 @@ module Metanorma
         content = CSV.parse_line(out).map do |x|
           x.sub!(/^(["'])(.+)\1/, "\\2")
           m = /^(.*?)(:\d+)?$/.match(x)
-          %{<toc-xpath depth='#{m[2]&.sub(/:/, '') || 1}'>#{m[1]}</toc-xpath>}
+          %{<toc-xpath depth='#{m[2]&.sub(':', '') || 1}'>#{m[1]}</toc-xpath>}
         end.join
         "<toc>#{content}</toc>"
       end

data/lib/metanorma/standoc/version.rb

@@ -19,6 +19,6 @@ module Metanorma
   end
 
   module Standoc
-    VERSION = "2.5.0".freeze
+    VERSION = "2.5.2".freeze
   end
 end

data/metanorma-standoc.gemspec

@@ -28,10 +28,11 @@ Gem::Specification.new do |spec|
   end
   spec.required_ruby_version = Gem::Requirement.new(">= 2.7.0")
 
+  spec.add_dependency "addressable", "~> 2.8.0"
   spec.add_dependency "asciidoctor", "~> 2.0.0"
   spec.add_dependency "iev", "~> 0.3.0"
   spec.add_dependency "isodoc", "~> 2.6.0"
-  spec.add_dependency "metanorma", ">= 1.5.0"
+  spec.add_dependency "metanorma", ">= 1.6.0"
   spec.add_dependency "metanorma-plugin-datastruct", "~> 0.2.0"
   spec.add_dependency "metanorma-plugin-glossarist", "~> 0.1.1"
   spec.add_dependency "metanorma-plugin-lutaml"

metadata

@@ -1,15 +1,29 @@
 --- !ruby/object:Gem::Specification
 name: metanorma-standoc
 version: !ruby/object:Gem::Version
-  version: 2.5.0
+  version: 2.5.2
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2023-08-07 00:00:00.000000000 Z
+date: 2023-08-10 00:00:00.000000000 Z
 dependencies:
+- !ruby/object:Gem::Dependency
+  name: addressable
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 2.8.0
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 2.8.0
 - !ruby/object:Gem::Dependency
   name: asciidoctor
   requirement: !ruby/object:Gem::Requirement
@@ -58,14 +72,14 @@ dependencies:
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.5.0
+        version: 1.6.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 1.5.0
+        version: 1.6.0
 - !ruby/object:Gem::Dependency
   name: metanorma-plugin-datastruct
   requirement: !ruby/object:Gem::Requirement