nokolexbor 0.3.4 → 0.3.6

data/lib/nokolexbor/node_set.rb

@@ -4,6 +4,11 @@ module Nokolexbor
   class NodeSet < Nokolexbor::Node
     include Enumerable
 
+    # Create a NodeSet with +document+ defaulting to +list+.
+    #
+    # @yield [Document]
+    #
+    # @return [Document]
     def self.new(document, list = [])
       obj = allocate
       obj.instance_variable_set(:@document, document)
@@ -12,6 +17,9 @@ module Nokolexbor
       obj
     end
 
+    # Iterate over each node.
+    #
+    # @yield [Node]
     def each
       return to_enum unless block_given?
 
@@ -21,6 +29,11 @@ module Nokolexbor
       self
     end
 
+    # Get the first +n+ elements of the NodeSet.
+    #
+    # @param n [Numeric,nil]
+    #
+    # @return [Node,Array<Node>] {Node} if +n+ is nil, otherwise {Array<Node>}
     def first(n = nil)
       return self[0] unless n
 
@@ -29,14 +42,19 @@ module Nokolexbor
       list
     end
 
+    # Get the last element of the NodeSet.
+    #
+    # @return [Node,nil]
     def last
       self[-1]
     end
 
+    # @return [Boolean] true if this NodeSet is empty.
     def empty?
       length == 0
     end
 
+    # @return [Integer] The index of the first node in this NodeSet that is equal to +node+ or meets the given block. Returns nil if no match is found.
     def index(node = nil)
       if node
         each_with_index { |member, j| return j if member == node }
@@ -46,6 +64,9 @@ module Nokolexbor
       nil
     end
 
+    # Get the content of all contained Nodes.
+    #
+    # @return [String]
     def content
       self.map(&:content).join
     end
@@ -54,10 +75,16 @@ module Nokolexbor
     alias_method :inner_text, :content
     alias_method :to_str, :content
 
+    # Get the inner html of all contained Nodes.
+    #
+    # @return [String]
     def inner_html(*args)
       self.map { |n| n.inner_html(*args) }.join
     end
 
+    # Convert this NodeSet to HTML.
+    #
+    # @return [String]
     def outer_html(*args)
       self.map { |n| n.outer_html(*args) }.join
     end
@@ -66,6 +93,9 @@ module Nokolexbor
     alias_method :to_html, :outer_html
     alias_method :serialize, :outer_html
 
+    # Remove all nodes in this NodeSet.
+    #
+    # @see Node#remove
     def remove
       self.each(&:remove)
     end
@@ -73,22 +103,32 @@ module Nokolexbor
     alias_method :unlink, :remove
     alias_method :to_ary, :to_a
 
+    # Destroy all nodes in the NodeSet.
+    #
+    # @see Node#destroy
     def destroy
       self.each(&:destroy)
     end
 
+    # @return [Node,nil] The last element of this NodeSet and removes it. Returns
+    #   +nil+ if the set is empty.
     def pop
       return nil if length == 0
 
       delete(last)
     end
 
+    # @return [Node,nil] The first element of this NodeSet and removes it. Returns
+    #   +nil+ if the set is empty.
     def shift
       return nil if length == 0
 
       delete(first)
     end
 
+    # @return [Boolean] true if two NodeSets contain the same number
+    #   of elements and each element is equal to the corresponding
+    #   element in the other NodeSet.
     def ==(other)
       return false unless other.is_a?(NodeSet)
       return false unless length == other.length
@@ -99,6 +139,8 @@ module Nokolexbor
       true
     end
 
+    # @return [NodeSet] A new NodeSet containing all the children of all the nodes in
+    #   the NodeSet.
     def children
       node_set = NodeSet.new(@document)
       each do |node|
@@ -107,6 +149,8 @@ module Nokolexbor
       node_set
     end
 
+    # @return [NodeSet] A new NodeSet containing all the nodes in the NodeSet
+    #   in reverse order.
     def reverse
       node_set = NodeSet.new(@document)
       (length - 1).downto(0) do |x|
@@ -115,6 +159,17 @@ module Nokolexbor
       node_set
     end
 
+    # Wrap all nodes of this NodeSet with +node_or_tags+.
+    #
+    # @see Node#wrap
+    #
+    # @return [NodeSet] +self+, to support chaining.
+    def wrap(node_or_tags)
+      map { |node| node.wrap(node_or_tags) }
+      self
+    end
+
+    # (see Node#xpath)
     def xpath(*args)
       paths, handler, ns, binds = extract_params(args)
 
@@ -127,6 +182,7 @@ module Nokolexbor
       end
     end
 
+    # (see Node#nokogiri_css)
     def nokogiri_css(*args)
       rules, handler, ns, _ = extract_params(args)
       paths = css_rules_to_xpath(rules, ns)

data/lib/nokolexbor/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Nokolexbor
-  VERSION = '0.3.4'
+  VERSION = '0.3.6'
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: nokolexbor
 version: !ruby/object:Gem::Version
-  version: 0.3.4
+  version: 0.3.6
 platform: ruby
 authors:
 - Yicheng Zhou
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2023-01-10 00:00:00.000000000 Z
+date: 2023-02-04 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake-compiler
@@ -259,28 +259,6 @@ files:
 - vendor/lexbor/source/lexbor/dom/interfaces/shadow_root.h
 - vendor/lexbor/source/lexbor/dom/interfaces/text.c
 - vendor/lexbor/source/lexbor/dom/interfaces/text.h
-- vendor/lexbor/source/lexbor/encoding/base.h
-- vendor/lexbor/source/lexbor/encoding/big5.c
-- vendor/lexbor/source/lexbor/encoding/config.cmake
-- vendor/lexbor/source/lexbor/encoding/const.h
-- vendor/lexbor/source/lexbor/encoding/decode.c
-- vendor/lexbor/source/lexbor/encoding/decode.h
-- vendor/lexbor/source/lexbor/encoding/encode.c
-- vendor/lexbor/source/lexbor/encoding/encode.h
-- vendor/lexbor/source/lexbor/encoding/encoding.c
-- vendor/lexbor/source/lexbor/encoding/encoding.h
-- vendor/lexbor/source/lexbor/encoding/euc_kr.c
-- vendor/lexbor/source/lexbor/encoding/gb18030.c
-- vendor/lexbor/source/lexbor/encoding/iso_2022_jp_katakana.c
-- vendor/lexbor/source/lexbor/encoding/jis0208.c
-- vendor/lexbor/source/lexbor/encoding/jis0212.c
-- vendor/lexbor/source/lexbor/encoding/multi.h
-- vendor/lexbor/source/lexbor/encoding/range.c
-- vendor/lexbor/source/lexbor/encoding/range.h
-- vendor/lexbor/source/lexbor/encoding/res.c
-- vendor/lexbor/source/lexbor/encoding/res.h
-- vendor/lexbor/source/lexbor/encoding/single.c
-- vendor/lexbor/source/lexbor/encoding/single.h
 - vendor/lexbor/source/lexbor/html/base.h
 - vendor/lexbor/source/lexbor/html/config.cmake
 - vendor/lexbor/source/lexbor/html/encoding.c

data/vendor/lexbor/source/lexbor/encoding/base.h

@@ -1,218 +0,0 @@
-/*
- * Copyright (C) 2019 Alexander Borisov
- *
- * Author: Alexander Borisov <borisov@lexbor.com>
- */
-
-#ifndef LEXBOR_ENCODING_BASE_H
-#define LEXBOR_ENCODING_BASE_H
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-#include "lexbor/core/base.h"
-#include "lexbor/encoding/const.h"
-
-
-#define LXB_ENCODING_VERSION_MAJOR 2
-#define LXB_ENCODING_VERSION_MINOR 0
-#define LXB_ENCODING_VERSION_PATCH 1
-
-#define LXB_ENCODING_VERSION_STRING                                            \
-        LEXBOR_STRINGIZE(LXB_ENCODING_VERSION_MAJOR) "."                       \
-        LEXBOR_STRINGIZE(LXB_ENCODING_VERSION_MINOR) "."                       \
-        LEXBOR_STRINGIZE(LXB_ENCODING_VERSION_PATCH)
-
-
-#define LXB_ENCODING_REPLACEMENT_BYTES ((lxb_char_t *) "\xEF\xBF\xBD")
-
-#define LXB_ENCODING_REPLACEMENT_BUFFER_LEN 1
-#define LXB_ENCODING_REPLACEMENT_BUFFER                                        \
-    (&(const lxb_codepoint_t) {LXB_ENCODING_REPLACEMENT_CODEPOINT})
-
-
-/*
- * In UTF-8 0x10FFFF value is maximum (inclusive)
- */
-enum {
-    LXB_ENCODING_REPLACEMENT_SIZE      = 0x03,
-    LXB_ENCODING_REPLACEMENT_CODEPOINT = 0xFFFD,
-    LXB_ENCODING_MAX_CODEPOINT         = 0x10FFFF,
-    LXB_ENCODING_ERROR_CODEPOINT       = 0x1FFFFF
-};
-
-enum {
-    LXB_ENCODING_ENCODE_OK           =  0x00,
-    LXB_ENCODING_ENCODE_ERROR        = -0x01,
-    LXB_ENCODING_ENCODE_SMALL_BUFFER = -0x02
-};
-
-enum {
-    LXB_ENCODING_DECODE_MAX_CODEPOINT = LXB_ENCODING_MAX_CODEPOINT,
-    LXB_ENCODING_DECODE_ERROR         = LXB_ENCODING_ERROR_CODEPOINT,
-    LXB_ENCODING_DECODE_CONTINUE      = 0x2FFFFF
-};
-
-enum {
-    LXB_ENCODING_DECODE_2022_JP_ASCII = 0x00,
-    LXB_ENCODING_DECODE_2022_JP_ROMAN,
-    LXB_ENCODING_DECODE_2022_JP_KATAKANA,
-    LXB_ENCODING_DECODE_2022_JP_LEAD,
-    LXB_ENCODING_DECODE_2022_JP_TRAIL,
-    LXB_ENCODING_DECODE_2022_JP_ESCAPE_START,
-    LXB_ENCODING_DECODE_2022_JP_ESCAPE,
-    LXB_ENCODING_DECODE_2022_JP_UNSET
-};
-
-enum {
-    LXB_ENCODING_ENCODE_2022_JP_ASCII = 0x00,
-    LXB_ENCODING_ENCODE_2022_JP_ROMAN,
-    LXB_ENCODING_ENCODE_2022_JP_JIS0208
-};
-
-typedef struct {
-    unsigned   need;
-    lxb_char_t lower;
-    lxb_char_t upper;
-}
-lxb_encoding_ctx_utf_8_t;
-
-typedef struct {
-    lxb_char_t first;
-    lxb_char_t second;
-    lxb_char_t third;
-}
-lxb_encoding_ctx_gb18030_t;
-
-typedef struct {
-    lxb_char_t lead;
-    bool       is_jis0212;
-}
-lxb_encoding_ctx_euc_jp_t;
-
-typedef struct {
-    lxb_char_t lead;
-    lxb_char_t prepand;
-    unsigned   state;
-    unsigned   out_state;
-    bool       out_flag;
-}
-lxb_encoding_ctx_2022_jp_t;
-
-typedef struct lxb_encoding_data lxb_encoding_data_t;
-
-typedef struct {
-    const lxb_encoding_data_t *encoding_data;
-
-    /* Out buffer */
-    lxb_codepoint_t           *buffer_out;
-    size_t                    buffer_length;
-    size_t                    buffer_used;
-
-    /*
-     * Bad code points will be replaced to user code point.
-     * If replace_to == 0 stop parsing and return error ot user.
-     */
-    const lxb_codepoint_t     *replace_to;
-    size_t                    replace_len;
-
-    /* Not for users */
-    lxb_codepoint_t           codepoint;
-    lxb_codepoint_t           second_codepoint;
-    bool                      prepend;
-    bool                      have_error;
-
-    lxb_status_t              status;
-
-    union {
-        lxb_encoding_ctx_utf_8_t   utf_8;
-        lxb_encoding_ctx_gb18030_t gb18030;
-        unsigned                   lead;
-        lxb_encoding_ctx_euc_jp_t  euc_jp;
-        lxb_encoding_ctx_2022_jp_t iso_2022_jp;
-    } u;
-}
-lxb_encoding_decode_t;
-
-typedef struct {
-    const lxb_encoding_data_t *encoding_data;
-
-    /* Out buffer */
-    lxb_char_t                *buffer_out;
-    size_t                    buffer_length;
-    size_t                    buffer_used;
-
-    /*
-     * Bad code points will be replaced to user bytes.
-     * If replace_to == NULL stop parsing and return error ot user.
-     */
-    const lxb_char_t          *replace_to;
-    size_t                    replace_len;
-
-    unsigned                  state;
-}
-lxb_encoding_encode_t;
-
-/*
-* Why can't I pass a char ** to a function which expects a const char **?
-* http://c-faq.com/ansi/constmismatch.html
-*
-* Short answer: use cast (const char **).
-*
-* For example:
-*     lxb_encoding_ctx_t ctx = {0};
-*     const lxb_encoding_data_t *enc;
-*
-*     lxb_char_t *data = (lxb_char_t *) "\x81\x30\x84\x36";
-*
-*     enc = lxb_encoding_data(LXB_ENCODING_GB18030);
-*
-*     enc->decode(&ctx, (const lxb_char_t **) &data, data + 4);
-*/
-typedef lxb_status_t
-(*lxb_encoding_encode_f)(lxb_encoding_encode_t *ctx, const lxb_codepoint_t **cp,
-                         const lxb_codepoint_t *end);
-
-typedef lxb_status_t
-(*lxb_encoding_decode_f)(lxb_encoding_decode_t *ctx,
-                         const lxb_char_t **data, const lxb_char_t *end);
-
-typedef int8_t
-(*lxb_encoding_encode_single_f)(lxb_encoding_encode_t *ctx, lxb_char_t **data,
-                                const lxb_char_t *end, lxb_codepoint_t cp);
-
-typedef lxb_codepoint_t
-(*lxb_encoding_decode_single_f)(lxb_encoding_decode_t *ctx,
-                                const lxb_char_t **data, const lxb_char_t *end);
-
-struct lxb_encoding_data {
-    lxb_encoding_t               encoding;
-    lxb_encoding_encode_f        encode;
-    lxb_encoding_decode_f        decode;
-    lxb_encoding_encode_single_f encode_single;
-    lxb_encoding_decode_single_f decode_single;
-    lxb_char_t                   *name;
-};
-
-typedef struct {
-    lxb_char_t      *name;
-    unsigned        size;
-    lxb_codepoint_t codepoint;
-}
-lxb_encoding_single_index_t;
-
-typedef lxb_encoding_single_index_t lxb_encoding_multi_index_t;
-
-typedef struct {
-    unsigned        index;
-    lxb_codepoint_t codepoint;
-}
-lxb_encoding_range_index_t;
-
-
-#ifdef __cplusplus
-} /* extern "C" */
-#endif
-
-#endif /* LEXBOR_ENCODING_BASE_H */