Merged in nimetu/ryzomcore/libxml2-html-parser (pull request #127)

implement html parser using libxml2
This commit is contained in:
Jan Boon 2015-04-16 17:01:47 +02:00
commit 8c006a46d4
4 changed files with 283 additions and 26 deletions

View file

@ -102,6 +102,9 @@ namespace NLGUI
// Browse // Browse
virtual void browse (const char *url); virtual void browse (const char *url);
// parse html string using libxml2 parser
virtual bool parseHtml(std::string htmlString);
// Refresh // Refresh
void refresh(); void refresh();
@ -199,6 +202,7 @@ namespace NLGUI
int luaBeginElement(CLuaState &ls); int luaBeginElement(CLuaState &ls);
int luaEndElement(CLuaState &ls); int luaEndElement(CLuaState &ls);
int luaShowDiv(CLuaState &ls); int luaShowDiv(CLuaState &ls);
int luaParseHtml(CLuaState &ls);
REFLECT_EXPORT_START(CGroupHTML, CGroupScrollText) REFLECT_EXPORT_START(CGroupHTML, CGroupScrollText)
REFLECT_LUA_METHOD("browse", luaBrowse) REFLECT_LUA_METHOD("browse", luaBrowse)
@ -210,6 +214,7 @@ namespace NLGUI
REFLECT_LUA_METHOD("beginElement", luaBeginElement) REFLECT_LUA_METHOD("beginElement", luaBeginElement)
REFLECT_LUA_METHOD("endElement", luaEndElement) REFLECT_LUA_METHOD("endElement", luaEndElement)
REFLECT_LUA_METHOD("showDiv", luaShowDiv) REFLECT_LUA_METHOD("showDiv", luaShowDiv)
REFLECT_LUA_METHOD("parseHtml", luaParseHtml)
REFLECT_STRING("url", getURL, setURL) REFLECT_STRING("url", getURL, setURL)
REFLECT_FLOAT("timeout", getTimeout, setTimeout) REFLECT_FLOAT("timeout", getTimeout, setTimeout)
REFLECT_EXPORT_END REFLECT_EXPORT_END
@ -228,7 +233,7 @@ namespace NLGUI
virtual void addText (const char * buf, int len); virtual void addText (const char * buf, int len);
// A link has been parsed // A link has been parsed
virtual void addLink (uint element_number, uint attribute_number, HTChildAnchor *anchor, const BOOL *present, const char **value); virtual void addLink (uint element_number, const BOOL *present, const char **value);
// A new begin HTML element has been parsed (<IMG> for exemple) // A new begin HTML element has been parsed (<IMG> for exemple)
virtual void beginElement (uint element_number, const BOOL *present, const char **value); virtual void beginElement (uint element_number, const BOOL *present, const char **value);
@ -251,6 +256,10 @@ namespace NLGUI
// the current request is terminated // the current request is terminated
virtual void requestTerminated(HTRequest *request); virtual void requestTerminated(HTRequest *request);
// libxml2 html parser functions
void htmlElement(xmlNode *node, int element_number);
void htmlWalkDOM(xmlNode *a_node);
// Get Home URL // Get Home URL
virtual std::string home(); virtual std::string home();
@ -668,6 +677,12 @@ namespace NLGUI
// read style attribute // read style attribute
void getStyleParams(const std::string &styleString, CStyleParams &style, bool inherit = true); void getStyleParams(const std::string &styleString, CStyleParams &style, bool inherit = true);
// load and render local html file (from bnp for example)
void doBrowseLocalFile(const std::string &filename);
// render html string as new browser page
bool renderHtmlString(const std::string &html);
private: private:
// decode all HTML entities // decode all HTML entities
static ucstring decodeHTMLEntities(const ucstring &str); static ucstring decodeHTMLEntities(const ucstring &str);
@ -699,6 +714,7 @@ namespace NLGUI
void checkImageDownload(); void checkImageDownload();
void addImageDownload(const std::string &url, CViewBase *img); void addImageDownload(const std::string &url, CViewBase *img);
std::string localImageName(const std::string &url); std::string localImageName(const std::string &url);
std::string getAbsoluteUrl(const std::string &url);
bool isTrustedDomain(const std::string &domain); bool isTrustedDomain(const std::string &domain);
void setImage(CViewBase *view, const std::string &file); void setImage(CViewBase *view, const std::string &file);

View file

@ -570,7 +570,7 @@ namespace NLGUI
// *************************************************************************** // ***************************************************************************
void CGroupHTML::addLink (uint element_number, uint /* attribute_number */, HTChildAnchor *anchor, const BOOL *present, const char **value) void CGroupHTML::addLink (uint element_number, const BOOL *present, const char **value)
{ {
if (_Browsing) if (_Browsing)
{ {
@ -591,16 +591,8 @@ namespace NLGUI
} }
else else
{ {
HTAnchor * dest = HTAnchor_followMainLink((HTAnchor *) anchor); // convert href from "?key=val" into "http://domain.com/?key=val"
if (dest) _Link.push_back(getAbsoluteUrl(suri));
{
C3WSmartPtr uri = HTAnchor_address(dest);
_Link.push_back ((const char*)uri);
}
else
{
_Link.push_back("");
}
} }
for(uint8 i = MY_HTML_A_ACCESSKEY; i < MY_HTML_A_Z_ACTION_SHORTCUT; i++) for(uint8 i = MY_HTML_A_ACCESSKEY; i < MY_HTML_A_Z_ACTION_SHORTCUT; i++)
@ -3832,6 +3824,10 @@ namespace NLGUI
stopBrowse (); stopBrowse ();
updateRefreshButton(); updateRefreshButton();
// Browsing
_Browsing = true;
updateRefreshButton();
// Home ? // Home ?
if (_URL == "home") if (_URL == "home")
_URL = home(); _URL = home();
@ -3851,17 +3847,24 @@ namespace NLGUI
_Connecting = true; _Connecting = true;
_ConnectingTimeout = ( times.thisFrameMs / 1000.0f ) + _TimeoutValue; _ConnectingTimeout = ( times.thisFrameMs / 1000.0f ) + _TimeoutValue;
// Save new url
_URL = finalUrl;
// file is probably from bnp (ingame help)
if (isLocal)
{
if (strlwr(finalUrl).find("file:/") == 0)
{
finalUrl = finalUrl.substr(6, finalUrl.size() - 6);
}
doBrowseLocalFile(finalUrl);
}
else
{
CButtonFreezer freezer; CButtonFreezer freezer;
this->visit(&freezer); this->visit(&freezer);
// Browsing
_Browsing = true;
updateRefreshButton();
// Save new url
_URL = finalUrl;
// display HTTP query // display HTTP query
//nlinfo("WEB: GET '%s'", finalUrl.c_str()); //nlinfo("WEB: GET '%s'", finalUrl.c_str());
@ -3877,12 +3880,7 @@ namespace NLGUI
C3WSmartPtr uri = HTParse(finalUrl.c_str(), NULL, PARSE_ALL); C3WSmartPtr uri = HTParse(finalUrl.c_str(), NULL, PARSE_ALL);
// Create an anchor // Create an anchor
#ifdef NL_OS_WINDOWS
if ((_LibWWW->Anchor = HTAnchor_findAddress(uri)) == NULL) if ((_LibWWW->Anchor = HTAnchor_findAddress(uri)) == NULL)
#else
// temporarily disable local URL's until LibWWW can be replaced.
if (isLocal || ((_LibWWW->Anchor = HTAnchor_findAddress(uri)) == NULL))
#endif
{ {
browseError((string("The page address is malformed : ")+(const char*)uri).c_str()); browseError((string("The page address is malformed : ")+(const char*)uri).c_str());
} }
@ -3919,6 +3917,8 @@ namespace NLGUI
} }
} }
} // !isLocal
_BrowseNextTime = false; _BrowseNextTime = false;
} }
@ -4107,6 +4107,57 @@ namespace NLGUI
#endif #endif
} }
// ***************************************************************************
void CGroupHTML::doBrowseLocalFile(const std::string &filename)
{
CIFile in;
if (in.open(filename))
{
std::string html;
while(!in.eof())
{
char buf[1024];
in.getline(buf, 1024);
html += std::string(buf) + "\n";
}
in.close();
if (!renderHtmlString(html))
{
browseError((string("Failed to parse html from file : ")+filename).c_str());
}
}
else
{
browseError((string("The page address is malformed : ")+filename).c_str());
}
}
// ***************************************************************************
bool CGroupHTML::renderHtmlString(const std::string &html)
{
bool success;
// clear content
beginBuild();
success = parseHtml(html);
// invalidate coords
endBuild();
// libwww would call requestTerminated() here
_Browsing = false;
if (_TitleString.empty())
{
setTitle(_TitlePrefix);
}
updateRefreshButton();
return success;
}
// *************************************************************************** // ***************************************************************************
void CGroupHTML::draw () void CGroupHTML::draw ()
@ -4516,7 +4567,7 @@ namespace NLGUI
beginElement(element_number, &present[0], &value[0]); beginElement(element_number, &present[0], &value[0]);
if (element_number == HTML_A) if (element_number == HTML_A)
addLink(element_number, 0, NULL, &present[0], &value[0]); addLink(element_number, &present[0], &value[0]);
return 0; return 0;
} }
@ -4650,6 +4701,15 @@ namespace NLGUI
return result; return result;
} }
// ***************************************************************************
std::string CGroupHTML::getAbsoluteUrl(const std::string &url)
{
if (HTURL_isAbsolute(url.c_str()))
return url;
return std::string(HTParse(url.c_str(), _URL.c_str(), PARSE_ALL));
}
// *************************************************************************** // ***************************************************************************
// CGroupHTML::CStyleParams style; // CGroupHTML::CStyleParams style;
// style.FontSize; // font-size: 10px; // style.FontSize; // font-size: 10px;

View file

@ -0,0 +1,178 @@
// Ryzom - MMORPG Framework <http://dev.ryzom.com/projects/ryzom/>
// Copyright (C) 2010 Winch Gate Property Limited
//
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU Affero General Public License as
// published by the Free Software Foundation, either version 3 of the
// License, or (at your option) any later version.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU Affero General Public License for more details.
//
// You should have received a copy of the GNU Affero General Public License
// along with this program. If not, see <http://www.gnu.org/licenses/>.
#include "stdpch.h"
#include <string>
#include <libxml/HTMLparser.h>
#include "nel/misc/types_nl.h"
#include "nel/gui/libwww.h"
#include "nel/gui/group_html.h"
#include "nel/gui/lua_ihm.h"
using namespace std;
using namespace NLMISC;
namespace NLGUI
{
// ***************************************************************************
void CGroupHTML::htmlElement(xmlNode *node, int element_number)
{
SGML_dtd *HTML_DTD = HTML_dtd ();
if (element_number < HTML_ELEMENTS)
{
CXMLAutoPtr ptr;
// load attributes into libwww structs
BOOL present[MAX_ATTRIBUTES];
const char *value[MAX_ATTRIBUTES];
std::string strvalues[MAX_ATTRIBUTES];
uint nbAttributes = std::min(MAX_ATTRIBUTES, HTML_DTD->tags[element_number].number_of_attributes);
for(uint i=0; i<nbAttributes; i++)
{
std::string name;
name = toLower(std::string(HTML_DTD->tags[element_number].attributes[i].name));
ptr = xmlGetProp(node, (const xmlChar *)name.c_str());
if (ptr)
{
// copy xmlChar to string (xmlChar will be released)
strvalues[i] = (const char *)(ptr);
// now use string pointer in value[] array
value[i] = strvalues[i].c_str();
present[i] = true;
}
else
{
value[i] = NULL;
present[i] = false;
}
}
if (element_number == HTML_A)
{
addLink(element_number, present, value);
}
beginElement(element_number, present, value);
}
else
{
beginUnparsedElement((const char *)(node->name), xmlStrlen(node->name));
}
// recursive - text content / child nodes
htmlWalkDOM(node->children);
// closing tag
if (element_number < HTML_ELEMENTS)
{
endElement(element_number);
}
else
{
endUnparsedElement((const char *)(node->name), xmlStrlen(node->name));
}
}
// ***************************************************************************
// recursive function to walk html document
void CGroupHTML::htmlWalkDOM(xmlNode *a_node)
{
SGML_dtd *HTML_DTD = HTML_dtd ();
uint element_number;
xmlNode *node = a_node;
while(node)
{
if (node->type == XML_TEXT_NODE)
{
addText((const char *)(node->content), xmlStrlen(node->content));
}
else
if (node->type == XML_ELEMENT_NODE)
{
// find libwww tag
for(element_number = 0; element_number<HTML_ELEMENTS; ++element_number)
{
if (xmlStrncasecmp(node->name, (const xmlChar *)HTML_DTD->tags[element_number].name, xmlStrlen(node->name)) == 0)
break;
}
htmlElement(node, element_number);
}
// move into next sibling
node = node->next;
}
}
// ***************************************************************************
bool CGroupHTML::parseHtml(std::string htmlString)
{
htmlParserCtxtPtr parser = htmlCreatePushParserCtxt(NULL, NULL, NULL, 0, NULL, XML_CHAR_ENCODING_NONE);
if (!parser)
{
nlwarning("Creating html parser context failed");
return false;
}
htmlCtxtUseOptions(parser, HTML_PARSE_NOBLANKS | HTML_PARSE_NOERROR | HTML_PARSE_NOWARNING | HTML_PARSE_NONET);
htmlParseChunk(parser, htmlString.c_str(), htmlString.size(), 0);
htmlParseChunk(parser, "", 0, 1);
bool success = true;
if (parser->myDoc)
{
xmlNode *root = xmlDocGetRootElement(parser->myDoc);
if (root)
{
htmlWalkDOM(root);
}
else
{
nlwarning("html root node failed");
success = false;
}
}
else
{
nlwarning("htmlstring parsing failed");
success = false;
}
htmlFreeParserCtxt(parser);
return success;
}
// ***************************************************************************
int CGroupHTML::luaParseHtml(CLuaState &ls)
{
const char *funcName = "parseHtml";
CLuaIHM::checkArgCount(ls, funcName, 1);
CLuaIHM::checkArgType(ls, funcName, 1, LUA_TSTRING);
std::string html = ls.toString(1);
parseHtml(html);
return 0;
}
}

View file

@ -322,7 +322,10 @@ namespace NLGUI
const char ** value) const char ** value)
{ {
// Do the work in the class // Do the work in the class
me->Parent->addLink (element_number, attribute_number, anchor, present, value); if (element_number == HTML_A)
{
me->Parent->addLink (element_number, present, value);
}
} }
// *************************************************************************** // ***************************************************************************