[htmlcxx] Added htmlcxx parser for source code displaying

It is compilable only together with spidermonkey, and only with meson
for now.

config2.h.in

@@ -100,6 +100,9 @@
 /* Define if you want: zlib support */
 #mesondefine CONFIG_GZIP
 
+/* Define if you want: htmlcxx support */
+#mesondefine CONFIG_HTMLCXX
+
 /* Define if you want: HTML highlighting support */
 #mesondefine CONFIG_HTML_HIGHLIGHT
 

meson.build

@@ -87,6 +87,7 @@ conf_data.set('CONFIG_LIBEV', get_option('libev'))
 conf_data.set('CONFIG_LIBEVENT', get_option('libevent'))
 conf_data.set('CONFIG_X', get_option('x'))
 conf_data.set('CONFIG_LIBDOM', get_option('libdom'))
+conf_data.set('CONFIG_HTMLCXX', get_option('htmlcxx'))
 
 #CONFIG_BOOKMARKS=true
 #CONFIG_XBEL_BOOKMARKS=true
@@ -156,7 +157,7 @@ conf_data.set10('HAVE_SYS_TIME_H', 1)
 
 compiler = meson.get_compiler('c')
 
-if conf_data.get('CONFIG_ECMASCRIPT')
+if conf_data.get('CONFIG_ECMASCRIPT') or conf_data.get('CONFIG_HTMLCXX')
     extracflags = ['-xc++', '-fpermissive']
 else
     extracflags = []
@@ -295,6 +296,11 @@ if conf_data.get('CONFIG_LIBDOM')
     deps += domdeps
 endif
 
+if conf_data.get('CONFIG_HTMLCXX')
+    htmlcxxdeps = dependency('htmlcxx')
+    deps += htmlcxxdeps
+endif
+
 if conf_data.get('CONFIG_SCRIPTING_PYTHON')
     python3deps = dependency('python3-embed')
     deps += python3deps

meson_options.txt

@@ -61,3 +61,4 @@ option('libev', type: 'boolean', value: false, description: 'compile with libev
 option('libevent', type: 'boolean', value: false, description: 'compile with libevent. Note that libev has precedence')
 option('x', type: 'boolean', value: false, description: 'use the X Window System')
 option('libdom', type: 'boolean', value: false, description: 'libdom')
+option('htmlcxx', type: 'boolean', value: false, description: 'htmlcxx')

src/document/htmlcxx/meson.build

@@ -0,0 +1 @@
+srcs += files('renderer.c')

src/document/htmlcxx/renderer.c

@@ -0,0 +1,86 @@
+/* Plain text document renderer */
+
+#ifdef HAVE_CONFIG_H
+#include "config.h"
+#endif
+
+#include <ctype.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "elinks.h"
+
+#include "bookmarks/bookmarks.h"
+#include "cache/cache.h"
+#include "config/options.h"
+#include "document/docdata.h"
+#include "document/document.h"
+#include "document/format.h"
+#include "document/options.h"
+#include "document/htmlcxx/renderer.h"
+#include "document/plain/renderer.h"
+#include "document/renderer.h"
+#include "globhist/globhist.h"
+#include "intl/charsets.h"
+#include "protocol/protocol.h"
+#include "protocol/uri.h"
+#include "terminal/color.h"
+#include "terminal/draw.h"
+#include "util/color.h"
+#include "util/error.h"
+#include "util/memory.h"
+#include "util/string.h"
+
+#include <htmlcxx/html/ParserDom.h>
+using namespace htmlcxx;
+
+struct source_renderer {
+	struct string tmp_buffer;
+	struct string *source;
+	char *enc;
+};
+
+static void
+walk_tree(struct string *buf, tree<HTML::Node> const &dom)
+{
+	tree<HTML::Node>::iterator it = dom.begin();
+	add_to_string(buf, it->text().c_str());
+
+	for (tree<HTML::Node>::sibling_iterator childIt = dom.begin(it); childIt != dom.end(it); ++childIt)
+	{
+		walk_tree(buf, childIt);
+	}
+	add_to_string(buf, it->closingText().c_str());
+}
+
+static int
+htmlcxx_main(struct source_renderer *renderer)
+{
+	std::string html = renderer->source->source;
+	HTML::ParserDom parser;
+	tree<HTML::Node> dom = parser.parseTree(html);
+	walk_tree(&renderer->tmp_buffer, dom);
+
+	return 0;
+}
+
+void
+render_source_document_cxx(struct cache_entry *cached, struct document *document,
+		      struct string *buffer)
+{
+	struct source_renderer renderer;
+	char *head = empty_string_or_(cached->head);
+
+	(void)get_convert_table(head, document->options.cp,
+					  document->options.assume_cp,
+					  &document->cp,
+					  &document->cp_status,
+					  document->options.hard_assume);
+
+	init_string(&renderer.tmp_buffer);
+	renderer.source = buffer;
+	renderer.enc = get_cp_mime_name(document->cp);
+	htmlcxx_main(&renderer);
+	render_plain_document(cached, document, &renderer.tmp_buffer);
+	done_string(&renderer.tmp_buffer);
+}

src/document/htmlcxx/renderer.h

@@ -0,0 +1,18 @@
+#ifndef EL__DOCUMENT_HTMLCXX_RENDERER_H
+#define EL__DOCUMENT_HTMLCXX_RENDERER_H
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+struct cache_entry;
+struct document;
+struct string;
+
+void render_source_document_cxx(struct cache_entry *cached, struct document *document, struct string *buffer);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif

src/document/meson.build

@@ -7,6 +7,9 @@ endif
 if conf_data.get('CONFIG_LIBDOM')
 	subdir('libdom')
 endif
+if conf_data.get('CONFIG_HTMLCXX')
+	subdir('htmlcxx')
+endif
 subdir('html')
 subdir('plain')
 srcs += files('docdata.c', 'document.c', 'format.c', 'forms.c', 'options.c', 'refresh.c', 'renderer.c')

src/document/renderer.c

@@ -18,6 +18,7 @@
 #include "document/dom/renderer.h"
 #include "document/html/frames.h"
 #include "document/html/renderer.h"
+#include "document/htmlcxx/renderer.h"
 #include "document/libdom/renderer.h"
 #include "document/plain/renderer.h"
 #include "document/renderer.h"
@@ -256,6 +257,14 @@ render_encoded_document(struct cache_entry *cached, struct document *document)
 		render_source_document(cached, document, &buffer);
 	}
 	else
+#endif
+#ifdef CONFIG_HTMLCXX
+	if (document->options.plain && cached->content_type
+	    && (!c_strcasecmp("text/html", cached->content_type)
+	    || !c_strcasecmp("application/xhtml+xml", cached->content_type))) {
+		render_source_document_cxx(cached, document, &buffer);
+	}
+	else
 #endif
 	if (document->options.plain) {
 #ifdef CONFIG_DOM