nokolexbor 0.2.6 → 0.3.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 653bea6b73eb4e74e61b68905d53d21586694489955f086eb018b1f51d097335
-  data.tar.gz: c451ee40b314bd34c3caf308021ccc7936beedf10ef26c793de78162dbf69113
+  metadata.gz: 6bfa3062e8718581cd8aabef75f0084a6626bd90301c9db4957ee975215edec1
+  data.tar.gz: 243f5c217c85750f63bfc0dcb8af575c019d3f2fd920f41769c0affc8a219802
 SHA512:
-  metadata.gz: eab8000ec155e0a8dfbff42c6602fca82481400e18df59ddd3920652f593154ebff25299d688eb5cd66e4a80cc666c65652b605f7f65e39242936e66b283a462
-  data.tar.gz: ea17c1bdee44e18955d981e5280cd7e34703a7de39710ef4fd13b47709699fab36071b40dba5908e014eded45e880643fa475263590e50a852afd72cec43f8a5
+  metadata.gz: e66b390ef696ed5cafa480065e8ac8fc66ce98d86a7cb15f2f59d8acb7df21381942e5fe6f0244e3b9c7bf54787131c5241a49faeaba328eb49e65af0ac3216c
+  data.tar.gz: d266e2a7c49a202aec4112667807158b54878cc014b42fc32492fc0977c06db76b1a5ded827d9810f2e7966ba3e5068fe54ab89915a4c56e499181ddc190b6d7

data/ext/nokolexbor/CMakeLists.txt

@@ -1,8 +1,11 @@
-cmake_minimum_required(VERSION 3.0)
+cmake_minimum_required(VERSION 2.8.12)
 
-set(VERSION "2.11.0")
-
-project(libxml2 VERSION ${VERSION} LANGUAGES C)
+IF(CMAKE_VERSION VERSION_LESS "3.0")
+	project(libxml2)
+ELSE()
+	cmake_policy(SET CMP0048 NEW)
+	project(libxml2 VERSION "2.11.0")
+ENDIF()
 
 include(CheckFunctionExists)
 include(CheckIncludeFiles)

data/ext/nokolexbor/config.h.cmake.in

@@ -11,7 +11,9 @@
 #cmakedefine HAVE_PTHREAD_H 1
 
 /* Define to 1 if you have the `rand_r' function. */
+#ifndef _WIN32
 #cmakedefine HAVE_RAND_R 1
+#endif
 
 /* Define to 1 if you have the <stdint.h> header file. */
 #cmakedefine HAVE_STDINT_H 1

data/ext/nokolexbor/extconf.rb

@@ -1,21 +1,60 @@
 require 'mkmf'
 require 'timeout'
 
+if ENV["CC"]
+  RbConfig::CONFIG["CC"] = RbConfig::MAKEFILE_CONFIG["CC"] = ENV["CC"]
+end
+
+# From: https://stackoverflow.com/questions/2108727
+# Cross-platform way of finding an executable in the $PATH.
+#
+#   which('ruby') #=> /usr/bin/ruby
+def which(cmd)
+  exts = ENV['PATHEXT'] ? ENV['PATHEXT'].split(';') : ['']
+  ENV['PATH'].split(File::PATH_SEPARATOR).each do |path|
+    exts.each { |ext|
+      exe = File.join(path, "#{cmd}#{ext}")
+      return exe if File.executable? exe
+    }
+  end
+  return nil
+end
+
 cmake_flags = [ ENV["CMAKE_FLAGS"] ]
-cmake_flags << "-DLEXBOR_BUILD_TESTS_CPP=OFF"
-cmake_flags << "-DLEXBOR_BUILD_SHARED=OFF"
-cmake_flags << "-DLEXBOR_BUILD_STATIC=ON"
+cmake_flags << "-DCMAKE_TRY_COMPILE_TARGET_TYPE=STATIC_LIBRARY"
+# Set system name explicitly when cross-compiling
+cmake_flags << "-DCMAKE_SYSTEM_NAME=Windows -DWIN32=1" if Gem.win_platform?
+# On Windows, Ruby-DevKit is MSYS-based, so ensure to use MSYS Makefiles.
+cmake_flags << "-G \"MSYS Makefiles\"" if Gem.win_platform? && !ENV['NOKOLEXBOR_CROSS_COMPILE']
+
+if ENV['NOKOLEXBOR_CROSS_COMPILE']
+  # use the same toolchain for cross-compiling lexbor
+  ['CC', 'CXX'].each do |env|
+    if RbConfig::CONFIG[env]
+      ENV[env] = RbConfig::CONFIG[env]
+    end
+  end
+  {'RANLIB' => 'RANLIB', 'AR' => 'AR', 'LD' => 'LINKER'}.each do |env, cmake_opt|
+    if RbConfig::CONFIG[env]
+      cmake_flags << "-DCMAKE_#{cmake_opt}=#{which(RbConfig::CONFIG[env])}"
+    end
+  end
+end
+
+lexbor_cmake_flags = cmake_flags + ["-DLEXBOR_BUILD_TESTS_CPP=OFF"]
+lexbor_cmake_flags << "-DLEXBOR_BUILD_SHARED=OFF"
+lexbor_cmake_flags << "-DLEXBOR_BUILD_STATIC=ON"
 
 if ENV['NOKOLEXBOR_DEBUG'] || ENV['NOKOLEXBOR_ASAN']
   CONFIG["optflags"] = "-O0"
   CONFIG["debugflags"] = "-ggdb3"
-  cmake_flags << "-DLEXBOR_OPTIMIZATION_LEVEL='-O0 -g'"
+  lexbor_cmake_flags << "-DLEXBOR_OPTIMIZATION_LEVEL='-O0 -g'"
 end
 
 if ENV['NOKOLEXBOR_ASAN']
   $LDFLAGS << " -fsanitize=address"
   $CFLAGS << " -fsanitize=address -DNOKOLEXBOR_ASAN"
-  cmake_flags << "-DLEXBOR_BUILD_WITH_ASAN=ON"
+  lexbor_cmake_flags << "-DLEXBOR_BUILD_WITH_ASAN=ON"
 end
 
 append_cflags("-DLEXBOR_STATIC")
@@ -35,7 +74,7 @@ end
 
 def self.run_cmake(timeout, args)
   # Set to process group so we can kill it and its children
-  pgroup = Gem.win_platform? ? :new_pgroup : :pgroup
+  pgroup = (Gem.win_platform? && !ENV['NOKOLEXBOR_CROSS_COMPILE']) ? :new_pgroup : :pgroup
   pid = Process.spawn("cmake #{args}", pgroup => true)
 
   Timeout.timeout(timeout) do
@@ -50,21 +89,6 @@ rescue Timeout::Error
   raise CMakeTimeout.new("cmake has exceeded its timeout of #{timeout}s")
 end
 
-# From: https://stackoverflow.com/questions/2108727
-# Cross-platform way of finding an executable in the $PATH.
-#
-#   which('ruby') #=> /usr/bin/ruby
-def which(cmd)
-  exts = ENV['PATHEXT'] ? ENV['PATHEXT'].split(';') : ['']
-  ENV['PATH'].split(File::PATH_SEPARATOR).each do |path|
-    exts.each { |ext|
-      exe = File.join(path, "#{cmd}#{ext}")
-      return exe if File.executable? exe
-    }
-  end
-  return nil
-end
-
 # From: https://github.com/flavorjones/mini_portile/blob/main/lib/mini_portile2/mini_portile.rb#L94
 def apply_patch(patch_file, chdir)
   case
@@ -109,9 +133,7 @@ Dir.chdir(LEXBOR_DIR) do
   Dir.mkdir("build") if !Dir.exist?("build")
 
   Dir.chdir("build") do
-    # On Windows, Ruby-DevKit is MSYS-based, so ensure to use MSYS Makefiles.
-    generator = "-G \"MSYS Makefiles\"" if Gem.win_platform?
-    run_cmake(10 * 60, ".. -DCMAKE_INSTALL_PREFIX:PATH=#{INSTALL_DIR} #{cmake_flags.join(' ')} #{generator}")
+    run_cmake(10 * 60, ".. -DCMAKE_INSTALL_PREFIX:PATH=#{INSTALL_DIR} #{lexbor_cmake_flags.join(' ')}")
     sys("#{MAKE} install")
   end
 end
@@ -121,7 +143,7 @@ Dir.chdir(EXT_DIR) do
   Dir.mkdir("build") if !Dir.exist?("build")
 
   Dir.chdir("build") do
-    run_cmake(10 * 60, "..#{Gem.win_platform? ? " -DLIBXML2_WITH_THREADS=OFF" : ""}")
+    run_cmake(10 * 60, ".. #{cmake_flags.join(' ')} #{Gem.win_platform? ? "-DLIBXML2_WITH_THREADS=OFF" : ""}")
   end
 end
 

data/ext/nokolexbor/memory.c

@@ -10,8 +10,8 @@
  * of more frequent GC.
  */
 
-#include <ruby.h>
 #include "lexbor/core/base.h"
+#include <ruby.h>
 
 // Disable using ruby memory functions when ASAN is enabled,
 // otherwise memory leak info will be all about ruby which
@@ -21,26 +21,26 @@
 void *
 lexbor_malloc(size_t size)
 {
-    return ruby_xmalloc(size);
+  return ruby_xmalloc(size);
 }
 
 void *
 lexbor_realloc(void *dst, size_t size)
 {
-    return ruby_xrealloc(dst, size);
+  return ruby_xrealloc(dst, size);
 }
 
 void *
 lexbor_calloc(size_t num, size_t size)
 {
-    return ruby_xcalloc(num, size);
+  return ruby_xcalloc(num, size);
 }
 
 void *
 lexbor_free(void *dst)
 {
-    ruby_xfree(dst);
-    return NULL;
+  ruby_xfree(dst);
+  return NULL;
 }
 
 #endif

data/ext/nokolexbor/nl_cdata.c

@@ -0,0 +1,44 @@
+#include "nokolexbor.h"
+
+VALUE cNokolexborCData;
+extern VALUE cNokolexborText;
+extern VALUE mNokolexbor;
+
+static VALUE
+nl_cdata_new(int argc, VALUE *argv, VALUE klass)
+{
+  lxb_dom_document_t *document;
+  VALUE rb_content;
+  VALUE rb_document;
+  VALUE rest;
+
+  rb_scan_args(argc, argv, "2*", &rb_content, &rb_document, &rest);
+
+  if (!rb_obj_is_kind_of(rb_document, cNokolexborDocument)) {
+    rb_raise(rb_eArgError, "Document must be a Nokolexbor::Document");
+  }
+
+  document = nl_rb_document_unwrap(rb_document);
+
+  const char* c_content = StringValuePtr(rb_content);
+  size_t content_len = RSTRING_LEN(rb_content);
+  lxb_dom_cdata_section_t *element = lxb_dom_document_create_cdata_section(document, (const lxb_char_t *)c_content, content_len);
+  if (element == NULL) {
+    rb_raise(rb_eRuntimeError, "Error creating text node");
+  }
+
+  VALUE rb_node = nl_rb_node_create(&element->text.char_data.node, rb_document);
+
+  if (rb_block_given_p()) {
+    rb_yield(rb_node);
+  }
+
+  return rb_node;
+}
+
+void Init_nl_cdata(void)
+{
+  cNokolexborCData = rb_define_class_under(mNokolexbor, "CDATA", cNokolexborText);
+
+  rb_define_singleton_method(cNokolexborCData, "new", nl_cdata_new, -1);
+}

data/ext/nokolexbor/nl_comment.c

@@ -0,0 +1,44 @@
+#include "nokolexbor.h"
+
+VALUE cNokolexborComment;
+extern VALUE cNokolexborCharacterData;
+extern VALUE mNokolexbor;
+
+static VALUE
+nl_comment_new(int argc, VALUE *argv, VALUE klass)
+{
+  lxb_dom_document_t *document;
+  VALUE rb_content;
+  VALUE rb_document;
+  VALUE rest;
+
+  rb_scan_args(argc, argv, "2*", &rb_content, &rb_document, &rest);
+
+  if (!rb_obj_is_kind_of(rb_document, cNokolexborDocument)) {
+    rb_raise(rb_eArgError, "Document must be a Nokolexbor::Document");
+  }
+
+  document = nl_rb_document_unwrap(rb_document);
+
+  const char* c_content = StringValuePtr(rb_content);
+  size_t content_len = RSTRING_LEN(rb_content);
+  lxb_dom_comment_t *element = lxb_dom_document_create_comment(document, (const lxb_char_t *)c_content, content_len);
+  if (element == NULL) {
+    rb_raise(rb_eRuntimeError, "Error creating comment");
+  }
+
+  VALUE rb_node = nl_rb_node_create(&element->char_data.node, rb_document);
+
+  if (rb_block_given_p()) {
+    rb_yield(rb_node);
+  }
+
+  return rb_node;
+}
+
+void Init_nl_comment(void)
+{
+  cNokolexborComment = rb_define_class_under(mNokolexbor, "Comment", cNokolexborCharacterData);
+
+  rb_define_singleton_method(cNokolexborComment, "new", nl_comment_new, -1);
+}

data/ext/nokolexbor/nl_document.c

@@ -26,12 +26,9 @@ nl_document_parse(VALUE self, VALUE rb_string_or_io)
 {
   VALUE id_read = rb_intern("read");
   VALUE rb_html;
-  if (rb_respond_to(rb_string_or_io, id_read))
-  {
+  if (rb_respond_to(rb_string_or_io, id_read)) {
     rb_html = rb_funcall(rb_string_or_io, id_read, 0);
-  }
-  else
-  {
+  } else {
     rb_html = rb_string_or_io;
   }
   const char *html_c = StringValuePtr(rb_html);
@@ -40,14 +37,12 @@ nl_document_parse(VALUE self, VALUE rb_string_or_io)
   lxb_html_document_t *document;
 
   document = lxb_html_document_create();
-  if (document == NULL)
-  {
+  if (document == NULL) {
     rb_raise(rb_eRuntimeError, "Error creating document");
   }
 
   lxb_status_t status = lxb_html_document_parse(document, (const lxb_char_t *)html_c, html_len);
-  if (status != LXB_STATUS_OK)
-  {
+  if (status != LXB_STATUS_OK) {
     nl_raise_lexbor_error(status);
   }
 
@@ -68,9 +63,28 @@ nl_rb_document_unwrap(VALUE rb_doc)
   return doc;
 }
 
+VALUE
+nl_document_get_title(VALUE rb_doc)
+{
+  size_t len;
+  lxb_char_t *str = lxb_html_document_title(nl_rb_document_unwrap(rb_doc), &len);
+  return str == NULL ? rb_str_new("", 0) : rb_utf8_str_new(str, len);
+}
+
+VALUE
+nl_document_set_title(VALUE rb_doc, VALUE rb_title)
+{
+  const char *c_title = StringValuePtr(rb_title);
+  size_t len = RSTRING_LEN(rb_title);
+  lxb_char_t *str = lxb_html_document_title_set(nl_rb_document_unwrap(rb_doc), (const lxb_char_t *)c_title, len);
+  return Qnil;
+}
+
 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);
+  rb_define_method(cNokolexborDocument, "title", nl_document_get_title, 0);
+  rb_define_method(cNokolexborDocument, "title=", nl_document_set_title, 1);
 }