nokolexbor 0.2.0 → 0.2.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 22cb30b169c61cb01d5a5b9d81a036ab3b47a99e34ff83f480bf1bd5b5a5eb0c
-  data.tar.gz: 12a185344aebf6e69088f65209c8750f9f946c5d62127e2f4c90ce9873a7368c
+  metadata.gz: 17230669785aa3a0bd2703cee73e72ccced0600d12fc52228125d4d6579b8290
+  data.tar.gz: 68fb22f67bcd9cfcad78eb3ecb2f860e71d72e78951987c3fae6ca8b03203c63
 SHA512:
-  metadata.gz: 663443bf82ab4061e2c15fdf00de6c762786125731346ed5afac99fbaabe26845e850510181e4d280294c9da9e36973a3cd3bc68a06d154b7563e1fe5000956b
-  data.tar.gz: a02a164cde0ce3209f518dd01d3d805a7c7010ef86e77fd398e136eb9d6ae60e1d4c913c7c0795fbde5f5ed57dffd824a9553a8df04122ca047ae11f84ea1df4
+  metadata.gz: 28b1fafc6889a6d020f7f7275ee9e668808334c6106fd79c68f514d233983df38169521d407726d6cc680d8e9f589b3914398bfdfc6a57a24fdda9a910741505
+  data.tar.gz: fd12b7be8b5f159680b652d2092144baab2fdd23e8b13c6d22cea69b6a9920019336cf07a5790148ff6fd1c7c0392d55989c2218abe876599a468397551dec9c

data/ext/nokolexbor/nl_document.c

@@ -41,11 +41,18 @@ nl_document_parse(VALUE self, VALUE rb_html)
     nl_raise_lexbor_error(status);
   }
 
-  return TypedData_Wrap_Struct(cNokolexborDocument, &nl_document_type, document);
+  return TypedData_Wrap_Struct(cNokolexborDocument, &nl_document_type, &document->dom_document);
+}
+
+static VALUE
+nl_document_new(VALUE self)
+{
+  return nl_document_parse(self, rb_str_new("", 0));
 }
 
 void Init_nl_document(void)
 {
   cNokolexborDocument = rb_define_class_under(mNokolexbor, "Document", cNokolexborNode);
+  rb_define_singleton_method(cNokolexborDocument, "new", nl_document_new, 0);
   rb_define_singleton_method(cNokolexborDocument, "parse", nl_document_parse, 1);
 }

data/ext/nokolexbor/nl_node.c

@@ -52,6 +52,40 @@ nl_rb_node_unwrap(VALUE rb_node)
   return node;
 }
 
+static VALUE
+nl_node_new(int argc, VALUE *argv, VALUE klass)
+{
+  lxb_dom_document_t *document;
+  lxb_dom_node_t *node;
+  VALUE rb_name;
+  VALUE rb_document;
+  VALUE rest;
+
+  rb_scan_args(argc, argv, "2*", &rb_name, &rb_document, &rest);
+
+  if (!rb_obj_is_kind_of(rb_document, cNokolexborDocument))
+  {
+    rb_raise(rb_eArgError, "Document must be a Nokolexbor::Document");
+  }
+
+  TypedData_Get_Struct(rb_document, lxb_dom_document_t, &nl_document_type, document);
+
+  lxb_dom_element_t *element = lxb_dom_document_create_element(document, StringValueCStr(rb_name), RSTRING_LEN(rb_name), NULL);
+  if (element == NULL)
+  {
+    rb_raise(rb_eRuntimeError, "Error creating element");
+  }
+
+  VALUE rb_node = nl_rb_node_create(&element->node, rb_document);
+
+  if (rb_block_given_p())
+  {
+    rb_yield(rb_node);
+  }
+
+  return rb_node;
+}
+
 static VALUE
 nl_node_content(VALUE self)
 {
@@ -251,7 +285,7 @@ mark_node_orders(lxb_dom_node_t *root)
 void sort_nodes_if_necessary(VALUE selector, lxb_dom_document_t *doc, lexbor_array_t *array)
 {
   // No need to sort if there's only one selector, the results are natually in document traversal order
-  if (strnstr(RSTRING_PTR(selector), ",", RSTRING_LEN(selector)) != NULL)
+  if (strstr(RSTRING_PTR(selector), ",") != NULL)
   {
     int need_order = 0;
     // Check if we have already markded orders, note that
@@ -574,6 +608,39 @@ nl_node_name(VALUE self)
   return rb_utf8_str_new(name, len);
 }
 
+static lxb_dom_node_t *
+nl_node_parse_fragment(lxb_html_document_t *doc, lxb_char_t *html, size_t size)
+{
+  size_t tag_name_len;
+  lxb_char_t *tag_name = lxb_tag_name_by_id(lxb_html_document_tags(doc), LXB_TAG__UNDEF, &tag_name_len);
+  if (tag_name == NULL)
+  {
+    rb_raise(rb_eRuntimeError, "Error getting tag name");
+  }
+  lxb_dom_element_t *element = lxb_dom_document_create_element(doc, tag_name, tag_name_len, NULL);
+  if (element == NULL)
+  {
+    rb_raise(rb_eRuntimeError, "Error creating element");
+  }
+  lxb_dom_node_t *frag_root = lxb_html_document_parse_fragment(doc, element, html, size);
+  if (frag_root == NULL)
+  {
+    rb_raise(rb_eArgError, "Error parsing HTML");
+  }
+  return frag_root;
+}
+
+static VALUE
+nl_node_fragment(VALUE self, VALUE html)
+{
+  Check_Type(html, T_STRING);
+  lxb_dom_node_t *node = nl_rb_node_unwrap(self);
+  lxb_dom_document_t *doc = node->owner_document;
+
+  lxb_dom_node_t *frag_root = nl_node_parse_fragment(doc, RSTRING_PTR(html), RSTRING_LEN(html));
+  return nl_rb_node_create(frag_root, nl_rb_document_get(self));
+}
+
 static VALUE
 nl_node_add_sibling(VALUE self, VALUE next_or_previous, VALUE new)
 {
@@ -596,22 +663,7 @@ nl_node_add_sibling(VALUE self, VALUE next_or_previous, VALUE new)
 
   if (TYPE(new) == T_STRING)
   {
-    size_t tag_name_len;
-    lxb_char_t *tag_name = lxb_tag_name_by_id(lxb_html_document_tags(doc), LXB_TAG__UNDEF, &tag_name_len);
-    if (tag_name == NULL)
-    {
-      rb_raise(rb_eRuntimeError, "Error getting tag name");
-    }
-    lxb_dom_element_t *element = lxb_dom_document_create_element(doc, tag_name, tag_name_len, NULL);
-    if (element == NULL)
-    {
-      rb_raise(rb_eRuntimeError, "Error creating element");
-    }
-    lxb_dom_node_t *frag_root = lxb_html_document_parse_fragment(doc, element, RSTRING_PTR(new), RSTRING_LEN(new));
-    if (frag_root == NULL)
-    {
-      rb_raise(rb_eArgError, "Error parsing HTML");
-    }
+    lxb_dom_node_t *frag_root = nl_node_parse_fragment(doc, RSTRING_PTR(new), RSTRING_LEN(new));
 
     while (frag_root->first_child != NULL)
     {
@@ -624,6 +676,7 @@ nl_node_add_sibling(VALUE self, VALUE next_or_previous, VALUE new)
   else if (rb_obj_class(new) == cNokolexborNode)
   {
     lxb_dom_node_t *node_new = nl_rb_node_unwrap(new);
+    lxb_dom_node_remove(node_new);
     insert_after ? lxb_dom_node_insert_after(node, node_new) : lxb_dom_node_insert_before(node, node_new);
   }
   else
@@ -641,22 +694,7 @@ nl_node_add_child(VALUE self, VALUE new)
 
   if (TYPE(new) == T_STRING)
   {
-    size_t tag_name_len;
-    lxb_char_t *tag_name = lxb_tag_name_by_id(lxb_html_document_tags(doc), LXB_TAG__UNDEF, &tag_name_len);
-    if (tag_name == NULL)
-    {
-      rb_raise(rb_eRuntimeError, "Error getting tag name");
-    }
-    lxb_dom_element_t *element = lxb_dom_document_create_element(doc, tag_name, tag_name_len, NULL);
-    if (element == NULL)
-    {
-      rb_raise(rb_eRuntimeError, "Error creating element");
-    }
-    lxb_dom_node_t *frag_root = lxb_html_document_parse_fragment(doc, element, RSTRING_PTR(new), RSTRING_LEN(new));
-    if (frag_root == NULL)
-    {
-      rb_raise(rb_eArgError, "Error parsing HTML");
-    }
+    lxb_dom_node_t *frag_root = nl_node_parse_fragment(doc, RSTRING_PTR(new), RSTRING_LEN(new));
 
     while (frag_root->first_child != NULL)
     {
@@ -747,9 +785,20 @@ nl_node_last_element_child(VALUE self)
   return Qnil;
 }
 
+static VALUE
+nl_node_clone(VALUE self)
+{
+  lxb_dom_node_t *node = nl_rb_node_unwrap(self);
+  lxb_dom_node_t *clone = lxb_dom_node_clone(node, 1);
+  return nl_rb_node_create(clone, nl_rb_document_get(self));
+}
+
 void Init_nl_node(void)
 {
   cNokolexborNode = rb_define_class_under(mNokolexbor, "Node", rb_cObject);
+  rb_undef_alloc_func(cNokolexborNode);
+
+  rb_define_singleton_method(cNokolexborNode, "new", nl_node_new, -1);
   rb_define_method(cNokolexborNode, "content", nl_node_content, 0);
   rb_define_method(cNokolexborNode, "[]", nl_node_get_attr, 1);
   rb_define_method(cNokolexborNode, "[]=", nl_node_set_attr, 2);
@@ -773,11 +822,13 @@ void Init_nl_node(void)
   rb_define_method(cNokolexborNode, "destroy", nl_node_destroy, 0);
   rb_define_method(cNokolexborNode, "attrs", nl_node_attrs, 0);
   rb_define_method(cNokolexborNode, "name", nl_node_name, 0);
+  rb_define_method(cNokolexborNode, "fragment", nl_node_fragment, 1);
   rb_define_method(cNokolexborNode, "add_sibling", nl_node_add_sibling, 2);
   rb_define_method(cNokolexborNode, "add_child", nl_node_add_child, 1);
   rb_define_method(cNokolexborNode, "node_type", nl_node_get_type, 0);
   rb_define_method(cNokolexborNode, "first_element_child", nl_node_first_element_child, 0);
   rb_define_method(cNokolexborNode, "last_element_child", nl_node_last_element_child, 0);
+  rb_define_method(cNokolexborNode, "clone", nl_node_clone, 0);
 
   rb_define_alias(cNokolexborNode, "attr", "[]");
   rb_define_alias(cNokolexborNode, "set_attr", "[]=");
@@ -787,4 +838,5 @@ void Init_nl_node(void)
   rb_define_alias(cNokolexborNode, "to_html", "outer_html");
   rb_define_alias(cNokolexborNode, "to_s", "outer_html");
   rb_define_alias(cNokolexborNode, "type", "node_type");
+  rb_define_alias(cNokolexborNode, "dup", "clone");
 }

data/ext/nokolexbor/xml_tree.c

@@ -131,7 +131,7 @@ xmlNodeGetContent(const lxb_dom_node_t *cur)
  * Get the root element of the document (doc->children is a list
  * containing possibly comments, PIs, etc ...).
  *
- * Returns the #xmlNodePtr for the root or NULL
+ * Returns the #lxb_dom_node_t_ptr for the root or NULL
  */
 lxb_dom_node_t_ptr
 xmlDocGetRootElement(const lxb_dom_document_t *doc) {
@@ -145,4 +145,16 @@ xmlDocGetRootElement(const lxb_dom_document_t *doc) {
         ret = ret->next;
     }
     return(ret);
+}
+
+/**
+ * xmlFreeNodeList:
+ * @cur:  the first node in the list
+ *
+ * Free a node and all its siblings, this is a recursive behaviour, all
+ * the children are freed too.
+ */
+void
+xmlFreeNodeList(lxb_dom_node_t_ptr cur) {
+    // Should never be called
 }

data/lib/nokolexbor/node.rb

@@ -67,6 +67,26 @@ module Nokolexbor
       end)
     end
 
+    def wrap(node)
+      case node
+      when String
+        new_parent = fragment(node).child
+      when Node
+        new_parent = node.dup
+      else
+        raise ArgumentError, "Requires a String or Node argument, and cannot accept a #{node.class}"
+      end
+
+      if parent
+        add_sibling(:next, new_parent)
+      else
+        new_parent.remove
+      end
+      new_parent.add_child(self)
+
+      self
+    end
+
     def matches?(selector)
       ancestors.last.css(selector).any? { |node| node == self }
     end

data/lib/nokolexbor/node_set.rb

@@ -4,10 +4,12 @@ module Nokolexbor
   class NodeSet < Node
     include Enumerable
 
-    def initialize(document, list = [])
-      @document = document
-      list.each { |x| self << x }
-      yield self if block_given?
+    def self.new(document, list = [])
+      obj = allocate
+      obj.instance_variable_set(:@document, document)
+      list.each { |x| obj << x }
+      yield obj if block_given?
+      obj
     end
 
     def each

data/lib/nokolexbor/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Nokolexbor
-  VERSION = '0.2.0'
+  VERSION = '0.2.2'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: nokolexbor
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.2.2
 platform: ruby
 authors:
 - Yicheng Zhou
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2022-12-14 00:00:00.000000000 Z
+date: 2022-12-15 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake-compiler