webtranslateit-hpricot 1.0.0 → 1.0.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 23b27a1746fc1bbaf18618228aa7355bb333c6b6b7bef6ba3baa228c096307bc
-  data.tar.gz: 79c391d5c90b510b969557372e9c944fca144b3bc42241908f894ed20dcbea70
+  metadata.gz: c9e0fe82d6e535d851172d7b897f527ed83ad9e57e600bd5af6637cdac73b2d4
+  data.tar.gz: 91a84b8ceea81a0d204cac790a00ba6cae9f453dcd3a9360059f7cafd4e7f6f2
 SHA512:
-  metadata.gz: 269e7a6e9f389c3045a5ee1f5eb824a5e99db52896b60b210d55ae51b66ac5894cebb2b3926837c3303ee6aec841289730f7e9e0ebfd653e06e94ecc0355f9ab
-  data.tar.gz: f0ac2d5cb9d96a7d2929e73d2ca94236fa067f1bd13832fb636eaad4638e73cac614f25ee5891cdf6bc8153545a8efea032994437adc7f065a41376e32ef1e96
+  metadata.gz: d45f0c8e2cb8fb1926e761ade66fb939da8574608d5e00a6929b29001f902fe37b8a59b159708fb6dfaf194538298d242a7454e719902b85350fa7af54e6dd8c
+  data.tar.gz: 3609483dd4ee0e5e3366d09c15b800b4f6fc09cbcee8f59ad3b00d290c70f68d0456bb26211bdc9764dd721c0ea9107f14cfb552385f3288a5f3d0432b7e97e6

data/CHANGELOG

@@ -1,3 +1,21 @@
+= 1.0.1
+=== 9 September 2026
+
+* Added a :void_elements parse option, naming which tag names are void instead
+  of taking HTML's whole list the way :html_void does. Void-ness belongs to the
+  document type rather than to the tag name: an XML vocabulary is free to reuse
+  a name HTML treats as void and mean a container by it. TYPO3's locallang XML
+  has a <meta> holding <type> and <description>, and html_void: true drops both
+  its children and its end tag. void_elements: %w[br] gives such a document the
+  one tag it needs and leaves every other name an ordinary element.
+
+  The set is matched exactly, since XML is case sensitive and this option is for
+  XML mode. A String or Symbol is accepted as a one-element set. An explicit
+  empty list means no element is void, and an explicit set overrides
+  :html_void, so a caller can narrow the list without restating it as a
+  negation. Serialization is unchanged: a void element is still written the way
+  the source wrote it, and HTML mode is untouched.
+
 = 1.0.0
 === 9 September 2026
 

data/hpricot.gemspec

@@ -1,6 +1,6 @@
 Gem::Specification.new do |s|
   s.name = 'webtranslateit-hpricot'
-  s.version = '1.0.0'
+  s.version = '1.0.1'
 
   s.authors = ['why the lucky stiff', 'WebTranslateIt']
   s.email = 'support@webtranslateit.com'

data/lib/hpricot/scanner.rb

@@ -17,13 +17,15 @@ module Hpricot
   #   * Every byte of the input appears in exactly one token's `raw`.
   #   * Attribute values are stored undecoded.
   class Scanner
-    def initialize(source, xml: false, fixup_tags: false, xhtml_strict: false, html_void: false)
+    def initialize(source, xml: false, fixup_tags: false, xhtml_strict: false, html_void: false,
+                   void_elements: nil)
       @xml = xml
       @fixup_tags = fixup_tags
       @xhtml_strict = xhtml_strict
       # Consumed by TreeBuilder; accepted here so Hpricot.scan can pass one
       # option hash to both.
       @html_void = html_void
+      @void_elements = void_elements
 
       # Lowest position from which no '>' (respectively ']') remains in the
       # document. Both start unknown. See gt_ahead?.

data/lib/hpricot/tree_builder.rb

@@ -9,10 +9,15 @@ module Hpricot
   # implicit closing driven by ElementContent (see #close_implied), matching
   # ext/hpricot_scan/hpricot_scan.rl:316-400.
   class TreeBuilder
-    def initialize(tokens, xml: false, fixup_tags: false, xhtml_strict: false, html_void: false)
+    def initialize(tokens, xml: false, fixup_tags: false, xhtml_strict: false, html_void: false,
+                   void_elements: nil)
       @tokens = tokens
       @xml = xml
-      @html_void = html_void
+      # Which tag names are void in XML mode, or nil for none (ordinary XML).
+      # :void_elements names the set outright; :html_void is shorthand for
+      # HTML's own list. An explicit set wins, so a caller can narrow the list
+      # without restating it as a negation.
+      @void_tags = void_elements || (html_void ? VOID_TAGS : nil)
       @fixup_tags = fixup_tags
       @xhtml_strict = xhtml_strict
     end
@@ -170,7 +175,15 @@ module Hpricot
         # XHTML way, with a self-closing slash". HTML mode keeps :EMPTY and so
         # keeps emitting '<br />'; html_void writes '<br>' as the source did,
         # which matters when the element sits inside translatable text.
-        e.allowed = :VOID if @html_void && VOID_TAGS[tok.name]
+        #
+        # void_elements names the set instead of taking HTML's whole list,
+        # because void-ness belongs to the document type rather than to the tag
+        # name. An XML vocabulary is free to reuse a name HTML happens to treat
+        # as void and mean a container by it -- TYPO3's locallang XML has a
+        # <meta> holding <type> and <description> -- and applying HTML's list
+        # there drops the children and the end tag. Such a caller asks for the
+        # one or two names it actually needs, typically %w[br].
+        e.allowed = :VOID if @void_tags&.key?(tok.name)
       else
         e.allowed = ElementContent[tok.name]
       end

data/lib/hpricot.rb

@@ -57,9 +57,26 @@ module Hpricot
     kw = { xml: !!opts[:xml],
            fixup_tags: !!opts[:fixup_tags],
            xhtml_strict: !!opts[:xhtml_strict],
-           html_void: !!opts[:html_void] }
+           html_void: !!opts[:html_void],
+           void_elements: void_element_set(opts[:void_elements]) }
     doc = TreeBuilder.new(Scanner.new(source, **kw).tokens, **kw).document
     doc.instance_variable_set(:@options, opts)
     doc
   end
+
+  # Normalises the :void_elements option into a name => true lookup.
+  #
+  # nil means the option was not given, and TreeBuilder falls back to
+  # :html_void. An explicit empty list is a different statement -- "no element
+  # is void" -- so it normalises to an empty Hash, which is truthy and
+  # therefore overrides :html_void the way any other explicit set does.
+  #
+  # Names are matched exactly. XML is case sensitive, and this option exists
+  # for XML mode, so a set naming 'br' does not cover '<BR>'.
+  def self.void_element_set(names)
+    return nil if names.nil?
+
+    Array(names).each_with_object({}) { |name, set| set[name.to_s] = true }.freeze
+  end
+  private_class_method :void_element_set
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: webtranslateit-hpricot
 version: !ruby/object:Gem::Version
-  version: 1.0.0
+  version: 1.0.1
 platform: ruby
 authors:
 - why the lucky stiff