nokogumbo 1.4.9 → 1.5.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
-SHA1:
-  metadata.gz: ecb515ec9c492a01dce23eed6eb7336695f166d6
-  data.tar.gz: c1ac4272ee113e56ecd72302847886042165a1fb
+SHA256:
+  metadata.gz: 96fa61565f78d5491e0b6d5b505cf936524745eb848b8b6584fc15e20c7ae35b
+  data.tar.gz: e5416f71bbe90323f04b8aad4dc48b28947e43a9eb46f446f8ca1444f519a07b
 SHA512:
-  metadata.gz: b979502577d944afa23cf0db4307494c6e53c320c3efc87c8066174d8ae3eee8902f60164375869dd6bd154abb1f6f69c42db9fbb0bc7fbd86d1b2f3e33bafd3
-  data.tar.gz: abd153cd24e504dc82c80d38ad91c8dd3dae25e913d650d0c92a8823cb1638dcf99baa43eb40a6a36573f9de7d14d78e2d77921062e341f0e82d49290a2f7ad7
+  metadata.gz: 676bf3585d38cd4ad5c72b8b3afd4952e248c747683ae1072dd43f6ce1ccd279177e4d0c75a9821ed76d32806333128152231349d8d113ae5d81279580b13004
+  data.tar.gz: 3459078d96977399e75551c4a3ee5623091f48569984b771e540ec111125f5af91e39a8d78cbd3ce9280326b1b9395dc4a0b0d7f0a72294876682cb9fe35e3d9

data/README.md

@@ -89,5 +89,5 @@ Installation
 Related efforts
 ============
 
-* [ruby-gumbo](https://github.com/galdor/ruby-gumbo#readme) - a ruby binding
+* [ruby-gumbo](https://github.com/nevir/ruby-gumbo#readme) - a ruby binding
 for the Gumbo HTML5 parser.

data/ext/nokogumboc/extconf.rb

@@ -40,4 +40,21 @@ unless have_library('gumbo', 'gumbo_parse')
   end
 end
 
+# We use some Gumbo Internals, and not all distros ship the internal headers.
+header_typedefs = {
+  'error.h' => 'GumboErrorType',
+  'insertion_mode.h' => 'GumboInsertionMode',
+  'parser.h' => 'GumboParser',
+  'string_buffer.h' => 'GumboStringBuffer',
+  'token_type.h' => 'GumboTokenType',
+}
+
+header_typedefs.each_pair do |header, type|
+  unless find_type(type, header)
+    require 'fileutils'
+    FileUtils.cp Dir["#{rakehome}/gumbo-parser/src/#{header}"],
+      "#{rakehome}/ext/nokogumboc"
+  end
+end
+
 create_makefile('nokogumboc')

data/ext/nokogumboc/nokogumbo.c

@@ -19,10 +19,13 @@
 //
 
 #include <ruby.h>
-#include <gumbo.h>
+#include "gumbo.h"
+#include "error.h"
+#include "parser.h"
 
 // class constants
 static VALUE Document;
+static VALUE XMLSyntaxError;
 
 #ifdef NGLIB
 #include <nokogiri.h>
@@ -181,11 +184,14 @@ static xmlNodePtr walk_tree(xmlDocPtr document, GumboNode *node) {
 }
 
 // Parse a string using gumbo_parse into a Nokogiri document
-static VALUE parse(VALUE self, VALUE string) {
-  GumboOutput *output = gumbo_parse_with_options(
-    &kGumboDefaultOptions, RSTRING_PTR(string),
-    (size_t) RSTRING_LEN(string)
-  );
+static VALUE parse(VALUE self, VALUE string, VALUE max_parse_errors) {
+  GumboOptions options;
+  memcpy(&options, &kGumboDefaultOptions, sizeof options);
+  options.max_errors = NUM2INT(max_parse_errors);
+
+  const char *input = RSTRING_PTR(string);
+  size_t input_len = RSTRING_LEN(string);
+  GumboOutput *output = gumbo_parse_with_options(&options, input, input_len);
   xmlDocPtr doc = xmlNewDoc(CONST_CAST "1.0");
 #ifdef NGLIB
   doc->type = XML_HTML_DOCUMENT_NODE;
@@ -210,9 +216,49 @@ static VALUE parse(VALUE self, VALUE string) {
         xmlAddChild((xmlNodePtr)doc, node);
     }
   }
-  gumbo_destroy_output(&kGumboDefaultOptions, output);
 
-  return Nokogiri_wrap_xml_document(Document, doc);
+  VALUE rdoc = Nokogiri_wrap_xml_document(Document, doc);
+
+  // Add parse errors to rdoc.
+  if (output->errors.length) {
+    GumboVector *errors = &output->errors;
+    GumboParser parser = { ._options = &options };
+    GumboStringBuffer msg;
+    VALUE rerrors = rb_ary_new2(errors->length);
+
+    gumbo_string_buffer_init(&parser, &msg);
+    for (int i=0; i < errors->length; i++) {
+      GumboError *err = errors->data[i];
+      gumbo_string_buffer_clear(&parser, &msg);
+      // Work around bug in gumbo_caret_diagnostic_to_string.
+      // See https://github.com/google/gumbo-parser/pull/371
+      // The bug occurs when the error starts with a newline (unless it's the
+      // first character in the input--but that shouldn't cause an error in
+      // the first place.
+      if (*err->original_text == '\n' && err->original_text != input)
+        --err->original_text;
+      gumbo_caret_diagnostic_to_string(&parser, err, input, &msg);
+      VALUE err_str = rb_str_new(msg.data, msg.length);
+      VALUE syntax_error = rb_class_new_instance(1, &err_str, XMLSyntaxError);
+      rb_iv_set(syntax_error, "@domain", INT2NUM(1)); // XML_FROM_PARSER
+      rb_iv_set(syntax_error, "@code", INT2NUM(1));   // XML_ERR_INTERNAL_ERROR
+      rb_iv_set(syntax_error, "@level", INT2NUM(2));  // XML_ERR_ERROR
+      rb_iv_set(syntax_error, "@file", Qnil);
+      rb_iv_set(syntax_error, "@line", INT2NUM(err->position.line));
+      rb_iv_set(syntax_error, "@str1", Qnil);
+      rb_iv_set(syntax_error, "@str2", Qnil);
+      rb_iv_set(syntax_error, "@str3", Qnil);
+      rb_iv_set(syntax_error, "@int1", INT2NUM(err->type));
+      rb_iv_set(syntax_error, "@column", INT2NUM(err->position.column));
+      rb_ary_push(rerrors, syntax_error);
+    }
+    rb_iv_set(rdoc, "@errors", rerrors);
+    gumbo_string_buffer_destroy(&parser, &msg);
+  }
+
+  gumbo_destroy_output(&options, output);
+
+  return rdoc;
 }
 
 // Initialize the Nokogumbo class and fetch constants we will use later
@@ -224,10 +270,11 @@ void Init_nokogumboc() {
   VALUE Nokogiri = rb_const_get(rb_cObject, rb_intern("Nokogiri"));
   VALUE HTML = rb_const_get(Nokogiri, rb_intern("HTML"));
   Document = rb_const_get(HTML, rb_intern("Document"));
+  VALUE XML = rb_const_get(Nokogiri, rb_intern("XML"));
+  XMLSyntaxError = rb_const_get(XML, rb_intern("SyntaxError"));
 
 #ifndef NGLIB
   // more class constants
-  VALUE XML = rb_const_get(Nokogiri, rb_intern("XML"));
   Element = rb_const_get(XML, rb_intern("Element"));
   Text = rb_const_get(XML, rb_intern("Text"));
   CDATA = rb_const_get(XML, rb_intern("CDATA"));
@@ -244,5 +291,5 @@ void Init_nokogumboc() {
 
   // define Nokogumbo class with a singleton parse method
   VALUE Gumbo = rb_define_class("Nokogumbo", rb_cObject);
-  rb_define_singleton_method(Gumbo, "parse", parse, 1);
+  rb_define_singleton_method(Gumbo, "parse", parse, 2);
 }

data/lib/nokogumbo.rb

@@ -4,14 +4,14 @@ require 'nokogumboc'
 module Nokogiri
   # Parse an HTML document.  +string+ contains the document.  +string+
   # may also be an IO-like object.  Returns a +Nokogiri::HTML::Document+.
-  def self.HTML5(string)
-    Nokogiri::HTML5.parse(string)
+  def self.HTML5(*args)
+    Nokogiri::HTML5.parse(*args)
   end
 
   module HTML5
     # Parse an HTML document.  +string+ contains the document.  +string+
     # may also be an IO-like object.  Returns a +Nokogiri::HTML::Document+.
-    def self.parse(string)
+    def self.parse(string, options={})
       if string.respond_to? :read
         string = string.read
       end
@@ -21,7 +21,7 @@ module Nokogiri
         string = reencode(string)
       end
 
-      Nokogumbo.parse(string.to_s)
+      Nokogumbo.parse(string.to_s, options[:max_parse_errors] || 0)
     end
 
     # Fetch and parse a HTML document from the web, following redirects,
@@ -67,7 +67,7 @@ module Nokogiri
 
       case response
       when Net::HTTPSuccess
-        doc = parse(reencode(response.body, response['content-type']))
+        doc = parse(reencode(response.body, response['content-type']), options)
         doc.instance_variable_set('@response', response)
         doc.class.send(:attr_reader, :response)
         doc
@@ -83,8 +83,8 @@ module Nokogiri
     # while fragment is on the Gumbo TODO list, simulate it by doing
     # a full document parse and ignoring the parent <html>, <head>, and <body>
     # tags, and collecting up the children of each.
-    def self.fragment(string)
-      doc = parse(string)
+    def self.fragment(*args)
+      doc = parse(*args)
       fragment = Nokogiri::HTML::DocumentFragment.new(doc)
 
       if doc.children.length != 1 or doc.children.first.name != 'html'

data/test-nokogumbo.rb

@@ -78,7 +78,7 @@ class TestNokogumbo < Minitest::Test
   end
 
   def test_html5_doctype
-    doc = Nokogumbo.parse("<!DOCTYPE html><html></html>")
+    doc = Nokogiri::HTML5.parse("<!DOCTYPE html><html></html>")
     assert_match /<!DOCTYPE html>/, doc.to_html
   end
 
@@ -125,6 +125,44 @@ class TestNokogumbo < Minitest::Test
     assert_equal ["html", "comment", "html", "comment"], doc.children.map(&:name)
   end
 
+  def test_parse_errors
+    doc = Nokogiri::HTML5("<!DOCTYPE html><html><!-- -- --></a>", max_parse_errors: 10)
+    assert_equal doc.errors.length, 2
+    doc = Nokogiri::HTML5("<!DOCTYPE html><html>", max_parse_errors: 10)
+    assert_empty doc.errors
+  end
+
+  def test_max_parse_errors
+    # This document contains 2 parse errors, but we force limit to 1.
+    doc = Nokogiri::HTML5("<!DOCTYPE html><html><!-- -- --></a>", max_parse_errors: 1)
+    assert_equal 1, doc.errors.length
+    doc = Nokogiri::HTML5("<!DOCTYPE html><html>", max_parse_errors: 1)
+    assert_empty doc.errors
+  end
+
+  def test_default_max_parse_errors
+    # This document contains 200 parse errors, but default limit is 0.
+    doc = Nokogiri::HTML5("<!DOCTYPE html><html>" + "</p>" * 200)
+    assert_equal 0, doc.errors.length
+  end
+
+  def test_parse_fragment_errors
+    doc = Nokogiri::HTML5.fragment("<\r\n", max_parse_errors: 10)
+    refute_empty doc.errors
+  end
+
+  def test_fragment_max_parse_errors
+    # This fragment contains 3 parse errors, but we force limit to 1.
+    doc = Nokogiri::HTML5.fragment("<!-- -- --></a>", max_parse_errors: 1)
+    assert_equal 1, doc.errors.length
+  end
+
+  def test_fragment_default_max_parse_errors
+    # This fragment contains 201 parse errors, but default limit is 0.
+    doc = Nokogiri::HTML5.fragment("</p>" * 200)
+    assert_equal 0, doc.errors.length
+  end
+
 private
 
   def buffer

metadata

@@ -1,27 +1,27 @@
 --- !ruby/object:Gem::Specification
 name: nokogumbo
 version: !ruby/object:Gem::Version
-  version: 1.4.9
+  version: 1.5.0
 platform: ruby
 authors:
 - Sam Ruby
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-08-01 00:00:00.000000000 Z
+date: 2018-01-27 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - '>='
+    - - ">="
       - !ruby/object:Gem::Version
         version: '0'
 description: Nokogumbo allows a Ruby program to invoke the Gumbo HTML5 parser and
@@ -32,11 +32,10 @@ extensions:
 - ext/nokogumboc/extconf.rb
 extra_rdoc_files: []
 files:
-- ext/nokogumboc/extconf.rb
-- ext/nokogumboc/nokogumbo.c
-- lib/nokogumbo.rb
 - LICENSE.txt
 - README.md
+- ext/nokogumboc/extconf.rb
+- ext/nokogumboc/nokogumbo.c
 - gumbo-parser/src/attribute.c
 - gumbo-parser/src/attribute.h
 - gumbo-parser/src/char_ref.c
@@ -69,10 +68,11 @@ files:
 - gumbo-parser/src/vector.c
 - gumbo-parser/src/vector.h
 - gumbo-parser/visualc/include/strings.h
+- lib/nokogumbo.rb
 - test-nokogumbo.rb
 homepage: https://github.com/rubys/nokogumbo/#readme
 licenses:
-- Apache 2.0
+- Apache-2.0
 metadata: {}
 post_install_message: 
 rdoc_options: []
@@ -80,17 +80,17 @@ require_paths:
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
-  - - '>='
+  - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - '>='
+  - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.0.14.1
+rubygems_version: 2.7.4
 signing_key: 
 specification_version: 4
 summary: Nokogiri interface to the Gumbo HTML5 parser