openehr 2.3.1 → 2.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9d2c2f002014051bc668449dbd7a7c79c2a75d4664d702670e80a9307b43ffb4
-  data.tar.gz: 374eccbd5be4cb9b76de4563a4648e5e2f346828a1abbe3988c6280de4278ecb
+  metadata.gz: 2d4cd8f720ac22f26e89145ebfd23cb474b2ad3e900c97fd6ff01c20a90bf638
+  data.tar.gz: ac478e309b65a00d314d3bf425b8d3ce25a7f07dd196e76c223508eca7410619
 SHA512:
-  metadata.gz: 7f5d3125c74f02accf2a6149ddfd7cfb5d79040184b1668de32a5850bf35a1464c2d448d140e865aa9e132adf3d8c93aca0d857a10c2ce96f54af2180a973c17
-  data.tar.gz: f0707a9dc5ed67e5ba4277177c1b5b3fe8e053251ca04f06f0abf326038d430a669dd5988de8793d4486dcae087f44d11f139c23b89df2f9409ee6195c9738d4
+  metadata.gz: 4f72f3492c23d458ed840da441359318e03a58c4bc55ee2d35df060ff635f4ab98e9be5aafafd126d149d81dcc355620a5931f65b05ff6d9dee65666df371a62
+  data.tar.gz: d4ca02374caaf65f3d6a29087fa788f4a6c57c9415d5c5ac576f9e746d56f4a6d25c4ef95fa3b2b56d905823f2ff7c08cb6bdb2f2f92944ab04d3b4f0734b97a

data/lib/openehr/parser/opt_parser.rb

@@ -40,7 +40,9 @@ module OpenEHR
       OCCURRENCE_PATH = '/occurrences'
 
       def parse
-        @opt = Nokogiri::XML::Document.parse(File.open(@filename))
+        @opt = File.open(@filename) do |file|
+          Nokogiri::XML::Document.parse(file, options: SAFE_PARSE_OPTIONS)
+        end
         @opt.remove_namespaces!
 
         uid = build_uid

data/lib/openehr/parser/xml_archetype_parser.rb

@@ -30,7 +30,9 @@ module OpenEHR
 
       def doc
         @doc ||= begin
-          parsed = Nokogiri::XML::Document.parse(File.open(@filename, 'rb:bom|utf-8'))
+          parsed = File.open(@filename, 'rb:bom|utf-8') do |file|
+            Nokogiri::XML::Document.parse(file, options: SAFE_PARSE_OPTIONS)
+          end
           parsed.remove_namespaces!
           parsed
         end

data/lib/openehr/parser.rb

@@ -1,6 +1,34 @@
+require 'nokogiri'
+
 module OpenEHR
   module Parser
     class Base
+      # Explicit, safe Nokogiri ParseOptions for untrusted OPT/archetype XML.
+      # Deliberately lists RECOVER|NONET|BIG_LINES by name rather than
+      # referencing Nokogiri::XML::ParseOptions::DEFAULT_XML - the whole
+      # point is to not silently follow that constant if a future Nokogiri
+      # release (or a downstream app pinning an older one) changes it.
+      #
+      # - RECOVER: parse malformed-but-well-intentioned XML leniently
+      #   (matches today's implicit default; see
+      #   docs/design/xxe-safe-parse-options-plan.md for real-fixture
+      #   evidence that at least one downstream consumer OPT currently
+      #   depends on this).
+      # - NONET: forbid network access during parsing. Nokogiri's own docs:
+      #   "UNSAFE to unset this option" for untrusted input.
+      # - BIG_LINES: line numbers as `long int`, unrelated to safety.
+      # - NOENT (entity substitution) and DTDLOAD (external DTD subset
+      #   loading) are deliberately NOT set - both default off in Nokogiri
+      #   and both documented "UNSAFE to set...for untrusted documents".
+      #   This is what actually gates XXE (see the plan's attack-proof
+      #   section: enabling either makes local-file and external-DTD
+      #   entity resolution succeed).
+      SAFE_PARSE_OPTIONS = Nokogiri::XML::ParseOptions.new(
+        Nokogiri::XML::ParseOptions::RECOVER |
+        Nokogiri::XML::ParseOptions::NONET |
+        Nokogiri::XML::ParseOptions::BIG_LINES
+      )
+
       attr_reader :filename
 
       def initialize(filename)

data/lib/openehr/rm/factory.rb

@@ -20,6 +20,8 @@ module OpenEHR
         terminology_id: 'TERMINOLOGY_ID'
       }.freeze
 
+      KNOWN_DERIVED_CACHE_TYPES = %w[TIMEZONE UID VERSION_TREE_ID].freeze
+
       def initialize(cobject)
         @cobject = cobject
       end
@@ -80,6 +82,11 @@ module OpenEHR
                   "and #{key.inspect} is not a known non-polymorphic attribute (keys: #{value.keys.inspect})"
           end
           Factory.create(type, **value)
+        rescue NameError
+          return nil if KNOWN_DERIVED_CACHE_TYPES.include?(type)
+
+          warn "openehr factory: unknown type \"#{type}\" for attribute #{key.inspect}; ignoring"
+          nil
         end
       end
 
@@ -610,4 +617,3 @@ module OpenEHR
   end
 end
 
-

data/lib/openehr/rm/support/identification.rb

@@ -249,6 +249,7 @@ module OpenEHR
               self.objectid = UID.new(:value => $1)
               self.creating_system_id = UID.new(:value => $2)
               self.version_tree_id = VersionTreeID.new(:value => $3)
+              @value = value
               @root = @oid
               @extension = @creating_system_id.value + '::' + @version_tree_id.value
             else

data/lib/openehr/serializer/rm_json_serializer.rb

@@ -10,7 +10,22 @@ module OpenEHR
     # which would otherwise make every subtree cyclic). A defensive
     # visited-object guard covers any other unexpected back-reference.
     class RMJSONSerializer
-      EXCLUDED_IVARS = [:@parent].freeze
+      # Instance variables that value=-style setters derive and cache from a
+      # single canonical attribute (e.g. `value`), but which are not part of
+      # openEHR ITS-JSON canonical form and have no corresponding
+      # OpenEHR::RM::Factory::<Type>Factory to reconstruct them on read - see
+      # docs/design/fix-rmjson-roundtrip-plan.md for the full inventory and
+      # how each was found (DvDate/DvTime/DvDateTime's value= in
+      # lib/openehr/rm/data_types/quantity/date_time.rb; UIDBasedID/
+      # ObjectVersionID's value= in lib/openehr/rm/support/identification.rb).
+      EXCLUDED_IVARS = [
+        :@parent,
+        :@year, :@month, :@day,                      # DvDate, DvDateTime
+        :@hour, :@minute, :@second, :@fractional_second, # DvTime, DvDateTime
+        :@timezone,                                  # DvTime (String), DvDateTime (Timezone)
+        :@root, :@extension,                         # UIDBasedID, HierObjectID, ObjectVersionID
+        :@oid, :@creating_system_id, :@version_tree_id # ObjectVersionID
+      ].freeze
 
       def initialize(rm_instance)
         @rm_instance = rm_instance

data/lib/openehr/version.rb

@@ -1,3 +1,3 @@
 module OpenEHR
-  VERSION = "2.3.1"
+  VERSION = "2.4.0"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: openehr
 version: !ruby/object:Gem::Version
-  version: 2.3.1
+  version: 2.4.0
 platform: ruby
 authors:
 - Shinji KOBAYASHI