nokolexbor 0.4.0 → 0.4.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: f60315f588c46f8ae2927633d3e34ee8f3aa3f6fb97575851bc967f306d216a5
-  data.tar.gz: 8e71ffb858e63f59284b94eee695b90808875c9af388175937a4e04cfbff715c
+  metadata.gz: 0055757bc7b9a92e9a2d37235722ca6a6945fb438855880ce10055dfa10d0fcb
+  data.tar.gz: b42cb00bc57ac3db09e928ff0b4ae3b3f7fb950312ab64f2606436184785dcc0
 SHA512:
-  metadata.gz: 70c50097ea4a844399ee4c74e7d0413a833dbdfd878208a8b7054ad2d01c2d96c35897b62848b199555b675a409f897a30e9cf90d2dd8c3e74c2fdf8b501340f
-  data.tar.gz: 7c772720a861bde11b10c1ff23d4796738a58eaa07d32b43478975a4feaab6402ef5fe8819eb9081619b50b970a422a40b613ce405fa2940b54da4929e352bc8
+  metadata.gz: 5527c86fea7d9efbbea6bcc1a2fcaaad506057ccf2d22d689f0de143b6c142f9bd1e8b33686ed2be1d01c37c1b02879c35e392a0ca81489f7f9cb125e569bf33
+  data.tar.gz: 4af797ac972b1ddfaafceec1ded5de2541daa92a366568f768f63f978bf3215e9a03295f598f4e05eea253aa2af8098ff8de33431119e7aeecad8a223de0c6f4

data/ext/nokolexbor/nl_node.c

@@ -1053,20 +1053,8 @@ static VALUE
 nl_node_first_element_child(VALUE self)
 {
   lxb_dom_node_t *parent = nl_rb_node_unwrap(self);
-  lxb_dom_node_t *cur;
+  lxb_dom_node_t *cur = parent->first_child;
 
-  if (parent == NULL) {
-    return Qnil;
-  }
-  switch (parent->type) {
-  case LXB_DOM_NODE_TYPE_ELEMENT:
-  case LXB_DOM_NODE_TYPE_ENTITY:
-  case LXB_DOM_NODE_TYPE_DOCUMENT:
-    cur = parent->first_child;
-    break;
-  default:
-    return Qnil;
-  }
   while (cur != NULL) {
     if (cur->type == LXB_DOM_NODE_TYPE_ELEMENT) {
       return nl_rb_node_create(cur, nl_rb_document_get(self));
@@ -1083,20 +1071,8 @@ static VALUE
 nl_node_last_element_child(VALUE self)
 {
   lxb_dom_node_t *parent = nl_rb_node_unwrap(self);
-  lxb_dom_node_t *cur;
+  lxb_dom_node_t *cur = parent->last_child;
 
-  if (parent == NULL) {
-    return Qnil;
-  }
-  switch (parent->type) {
-  case LXB_DOM_NODE_TYPE_ELEMENT:
-  case LXB_DOM_NODE_TYPE_ENTITY:
-  case LXB_DOM_NODE_TYPE_DOCUMENT:
-    cur = parent->last_child;
-    break;
-  default:
-    return Qnil;
-  }
   while (cur != NULL) {
     if (cur->type == LXB_DOM_NODE_TYPE_ELEMENT) {
       return nl_rb_node_create(cur, nl_rb_document_get(self));
@@ -1155,6 +1131,16 @@ nl_node_inspect(int argc, VALUE *argv, VALUE self)
   return rb_call_super(argc, argv);
 }
 
+/**
+ * @return [Integer] The node's location at the source HTML. Returns 0 if the node is not parsed from a HTML string.
+ */
+static VALUE
+nl_node_source_location(VALUE self)
+{
+  lxb_dom_node_t *node = nl_rb_node_unwrap(self);
+  return ULONG2NUM(node->source_location);
+}
+
 void Init_nl_node(void)
 {
   cNokolexborNode = rb_define_class_under(mNokolexbor, "Node", rb_cObject);
@@ -1199,6 +1185,7 @@ void Init_nl_node(void)
   rb_define_method(cNokolexborNode, "last_element_child", nl_node_last_element_child, 0);
   rb_define_method(cNokolexborNode, "clone", nl_node_clone, 0);
   rb_define_method(cNokolexborNode, "inspect", nl_node_inspect, -1);
+  rb_define_method(cNokolexborNode, "source_location", nl_node_source_location, 0);
 
   rb_define_alias(cNokolexborNode, "attr", "[]");
   rb_define_alias(cNokolexborNode, "get_attribute", "[]");

data/lib/nokolexbor/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Nokolexbor
-  VERSION = '0.4.0'
+  VERSION = '0.4.1'
 end

data/patches/0005-lexbor-add-source-location-to-node.patch

@@ -0,0 +1,65 @@
+diff --git a/source/lexbor/dom/interfaces/node.h b/source/lexbor/dom/interfaces/node.h
+index acd0c1c..f436257 100755
+--- a/source/lexbor/dom/interfaces/node.h
++++ b/source/lexbor/dom/interfaces/node.h
+@@ -46,6 +46,8 @@ struct lxb_dom_node {
+     uintptr_t              prefix;     /* lowercase: lalala */
+     uintptr_t              ns;         /* namespace */
+ 
++    size_t                 source_location; /* char offset to the source HTML */
++
+     lxb_dom_document_t     *owner_document;
+ 
+     lxb_dom_node_t         *next;
+diff --git a/source/lexbor/html/tokenizer.c b/source/lexbor/html/tokenizer.c
+index 741bced..6343c65 100755
+--- a/source/lexbor/html/tokenizer.c
++++ b/source/lexbor/html/tokenizer.c
+@@ -309,6 +309,7 @@ lxb_html_tokenizer_chunk(lxb_html_tokenizer_t *tkz, const lxb_char_t *data,
+ 
+     tkz->is_eof = false;
+     tkz->status = LXB_STATUS_OK;
++    tkz->first = data;
+     tkz->last = end;
+ 
+     while (data < end) {
+diff --git a/source/lexbor/html/tokenizer.h b/source/lexbor/html/tokenizer.h
+index ba9602f..08d0d9a 100755
+--- a/source/lexbor/html/tokenizer.h
++++ b/source/lexbor/html/tokenizer.h
+@@ -72,6 +72,7 @@ struct lxb_html_tokenizer {
+     lxb_char_t                       *pos;
+     const lxb_char_t                 *end;
+     const lxb_char_t                 *begin;
++    const lxb_char_t                 *first;
+     const lxb_char_t                 *last;
+ 
+     /* Entities */
+diff --git a/source/lexbor/html/tree.c b/source/lexbor/html/tree.c
+index 8c42990..28c97cc 100755
+--- a/source/lexbor/html/tree.c
++++ b/source/lexbor/html/tree.c
+@@ -484,6 +484,7 @@ lxb_html_tree_append_attributes(lxb_html_tree_t *tree,
+ 
+         attr->node.local_name = token_attr->name->attr_id;
+         attr->node.ns = ns;
++        attr->node.source_location = token_attr->name_begin - tree->tkz_ref->first;
+ 
+         /* Fix for adjust MathML/SVG attributes */
+         if (tree->before_append_attr != NULL) {
+diff --git a/source/lexbor/html/tree.h b/source/lexbor/html/tree.h
+index 231239d..bc0249e 100755
+--- a/source/lexbor/html/tree.h
++++ b/source/lexbor/html/tree.h
+@@ -266,8 +266,10 @@ lxb_inline lxb_dom_node_t *
+ lxb_html_tree_create_node(lxb_html_tree_t *tree,
+                           lxb_tag_id_t tag_id, lxb_ns_id_t ns)
+ {
+-    return (lxb_dom_node_t *) lxb_html_interface_create(tree->document,
++    lxb_dom_node_t *node = (lxb_dom_node_t *) lxb_html_interface_create(tree->document,
+                                                         tag_id, ns);
++    node->source_location = (tree->tkz_ref->token->begin > tree->tkz_ref->first ? tree->tkz_ref->token->begin : tree->tkz_ref->begin) - tree->tkz_ref->first;
++    return node;
+ }
+ 
+ lxb_inline bool

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: nokolexbor
 version: !ruby/object:Gem::Version
-  version: 0.4.0
+  version: 0.4.1
 platform: ruby
 authors:
 - Yicheng Zhou
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2023-03-22 00:00:00.000000000 Z
+date: 2023-04-04 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake-compiler
@@ -131,6 +131,7 @@ files:
 - patches/0002-lexbor-match-id-class-case-sensitive.patch
 - patches/0003-lexbor-attach-template-content-to-self.patch
 - patches/0004-lexbor-fix-template-clone.patch
+- patches/0005-lexbor-add-source-location-to-node.patch
 - vendor/lexbor/CMakeLists.txt
 - vendor/lexbor/config.cmake
 - vendor/lexbor/feature.cmake