Better handle some cases & Release 1.0.2 (#55)

* Better handle mixed content in some cases
* Correctly set node content, even if content is blank
* Add specs to ensure compatibility with XML to JSON conventions
* Release 1.0.2

shard.yml

@@ -3,7 +3,7 @@ name: oq
 description: |
   A performant, and portable jq wrapper thats facilitates the consumption and output of formats other than JSON; using jq filters to transform the data.
 
-version: 1.0.1
+version: 1.0.2
 
 authors:
   - George Dietrich <[email protected]>

snap/snapcraft.yaml

@@ -1,5 +1,5 @@
 name: oq
-version: '1.0.1'
+version: '1.0.2'
 summary: A performant, and portable jq wrapper to support formats other than JSON
 description: |
   A performant, and portable jq wrapper thats facilitates the consumption and output of formats other than JSON; using jq filters to transform the data.

spec/converters/xml_spec.cr

@@ -141,11 +141,66 @@ XML_ALL_EMPTY = <<-XML
   <one> </one>
   <two>
   </two>
+  <three/>
+  <four></four>
 </root>
 XML
 
 describe OQ::Converters::Xml do
   describe ".deserialize" do
+    # See https://www.xml.com/pub/a/2006/05/31/converting-between-xml-and-json.html
+    describe "conventions" do
+      describe "an empty element" do
+        it "self closing" do
+          run_binary("<e/>", args: ["-i", "xml", "-c", "."]) do |output|
+            output.should eq %({"e":null}\n)
+          end
+        end
+
+        it "non self closing" do
+          run_binary("<e></e>", args: ["-i", "xml", "-c", "."]) do |output|
+            output.should eq %({"e":null}\n)
+          end
+        end
+      end
+
+      it "an element with pure text content" do
+        run_binary("<e>text</e>", args: ["-i", "xml", "-c", "."]) do |output|
+          output.should eq %({"e":"text"}\n)
+        end
+      end
+
+      it "an empty element with attributes" do
+        run_binary(%(<e name="value" />), args: ["-i", "xml", "-c", "."]) do |output|
+          output.should eq %({"e":{"@name":"value"}}\n)
+        end
+      end
+
+      it "an element with pure text content and attributes" do
+        run_binary(%(<e name="value">text</e>), args: ["-i", "xml", "-c", "."]) do |output|
+          output.should eq %({"e":{"@name":"value","#text":"text"}}\n)
+        end
+      end
+
+      it "an element containing elements with different names" do
+        run_binary(%(<e> <a>text</a> <b>text</b> </e>), args: ["-i", "xml", "-c", "."]) do |output|
+          output.should eq %({"e":{"a":"text","b":"text"}}\n)
+        end
+      end
+
+      it "an element containing elements with identical names" do
+        run_binary(%(<e> <a>text</a> <a>text</a> </e>), args: ["-i", "xml", "-c", "."]) do |output|
+          output.should eq %({"e":{"a":["text","text"]}}\n)
+        end
+      end
+
+      it "an element containing elements and contiguous text" do
+        run_binary(%(<e>text<a>text</a></e>), args: ["-i", "xml", "-c", "."]) do |output|
+          output.should eq %({"e":{"#text":"text","a":"text"}}\n)
+        end
+      end
+    end
+
     describe "should raise if invalid" do
       it "should output correctly" do
         run_binary(%(<root id="1<child/></root>), args: ["-i", "xml", "-c", "."]) do |_, status, error|
@@ -167,7 +222,7 @@ describe OQ::Converters::Xml do
       describe "that has only empty children elements" do
         it "should output an object with null values" do
           run_binary(XML_ALL_EMPTY, args: ["-i", "xml", "-c", "."]) do |output|
-            output.should eq %({"root":{"one":null,"two":null}}\n)
+            output.should eq %({"root":{"one":" ","two":"\\n  ","three":null,"four":null}}\n)
           end
         end
       end
@@ -221,9 +276,9 @@ describe OQ::Converters::Xml do
       end
 
       describe "with mixed content" do
-        it "should output correctly" do
+        it "with a single #text node" do
           run_binary(%(<root>x<y>z</y></root>), args: ["-i", "xml", "-c", ".root"]) do |output|
-            output.should eq %({"y":"z"}\n)
+            output.should eq %({"#text":"x","y":"z"}\n)
           end
         end
       end

src/converters/xml.cr

@@ -18,19 +18,16 @@ module OQ::Converters::Xml
           xml.read
         end
 
-        # TODO: clean up after crystal-lang/crystal#8186 is released
-        if node = xml.expand
-          process_element_node node, builder
-        else
-          raise XML::Error.new LibXML.xmlGetLastError
-        end
+        process_element_node xml.expand, builder
       end
     end
   end
 
   private def self.process_element_node(node : XML::Node, builder : JSON::Builder) : Nil
     # If the node doesn't have nested elements nor attributes; just emit a scalar value
-    return builder.field node.name, get_node_value node if !has_nested_elements(node) && node.attributes.empty?
+    if !has_nested_elements(node) && node.attributes.empty?
+      return builder.field node.name, get_node_value node
+    end
 
     # Otherwise process the node as a key/value pair
     builder.field node.name do
@@ -67,7 +64,14 @@ module OQ::Converters::Xml
     # Determine how to process a node's children
     node.children.group_by(&.name).each do |name, children|
       # Skip non significant whitespace; Skip mixed character input
-      next if children.first.text? && has_nested_elements(node)
+      if children.first.text? && has_nested_elements(node)
+        # Only emit text content if there is only one child
+        if children.size == 1
+          builder.field "#text", children.first.content
+        end
+
+        next
+      end
 
       # Array
       if children.size > 1
@@ -89,12 +93,12 @@ module OQ::Converters::Xml
   end
 
   private def self.get_node_value(node : XML::Node) : String?
-    node.children.empty? || node.children.first.content.blank? ? nil : node.children.first.content
+    node.children.empty? ? nil : node.children.first.content
   end
 
   def self.serialize(input : IO, output : IO, **args) : Nil
-    json = JSON::PullParser.new(input)
-    builder = XML::Builder.new(output)
+    json = JSON::PullParser.new input
+    builder = XML::Builder.new output
     indent, prolog, root, xml_item = self.parse_args(args)
 
     builder.indent = indent

src/oq.cr

@@ -6,7 +6,7 @@ require "./converters/*"
 
 # A performant, and portable jq wrapper thats facilitates the consumption and output of formats other than JSON; using jq filters to transform the data.
 module OQ
-  VERSION = "1.0.1"
+  VERSION = "1.0.2"
 
   # The support formats that can be converted to/from.
   enum Format