C++ XmlDoc::printDocForProCog方法代码示例

bool gotXmlDoc ( void *state ) {
	// cast it
	State8 *st = (State8 *)state;
	// get the xmldoc
	XmlDoc *xd = &st->m_xd;

	// if we loaded from old title rec, it should be there!


	// . save the ips.txt file if we are the test coll
	// . saveTestBuf() is a function in Msge1.cpp
	//if ( xd && xd->m_coll && ! strcmp ( xd->m_coll , "qatest123")) 
	//	// use same dir that XmlDoc::getTestDir() would use
	//	saveTestBuf ( "test-page-parser" );

	// error?
	if ( g_errno ) return sendErrorReply ( st , g_errno );

	// shortcut
	SafeBuf *xbuf = &st->m_xbuf;

	bool printIt = false;
	if ( st->m_u && st->m_u[0] ) printIt = true;
	if ( st->m_docId != -1LL ) printIt = true;
	if ( st->m_donePrinting ) printIt = false;

	// do not re-call this if printDocForProCog blocked... (check length())
	if ( printIt ) {
		// mark as done
		st->m_donePrinting = true;
		// always re-compute the page inlinks dynamically, do not
		// use the ptr_linkInfo1 stored in titlerec!!
		// NO! not if set from titlerec/docid
		if ( st->m_recompute )
			xd->m_linkInfo1Valid = false;
		// try a recompute regardless, because we do not store the
		// bad inlinkers, and ppl want to see why they are bad!
		//xd->m_linkInfo1Valid = false;
		// now get the meta list, in the process it will print out a 
		// bunch of junk into st->m_xbuf
		//char *metalist = xd->getMetaList ( );
		//if ( ! metalist ) return sendErrorReply ( st , g_errno );
		// return false if it blocked
		//if ( metalist == (void *)-1 ) return false;
		// for debug...
		//if ( ! xd->m_indexCode ) xd->doConsistencyTest ( false );
		// . print it out
		// . returns false if blocks, true otherwise
		// . sets g_errno on error
		if ( ! xd->printDocForProCog ( xbuf , &st->m_r ) )
			return false;
		// error?
		if ( g_errno ) return sendErrorReply ( st , g_errno );
	}

	long isXml = st->m_r.getLong("xml",0);
	char ctype2 = CT_HTML;
	if ( isXml ) ctype2 = CT_XML;
	// now encapsulate it in html head/tail and send it off
	bool status = g_httpServer.sendDynamicPage( st->m_s , 
						    xbuf->getBufStart(), 
						    xbuf->length() ,
						    -1, //cachtime
						    false ,//postreply?
						    &ctype2,
						    -1 , //httpstatus
						    NULL,//cookie
						    "utf-8");
	// delete the state now
	if ( st->m_freeIt ) {
		mdelete ( st , sizeof(State8) , "PageParser" );
		delete (st);
	}
	// return the status
	return status;
}

// returns false if blocked, true otherwise
bool processLoop ( void *state ) {
	// get it
	State2 *st = (State2 *)state;
	// get the tcp socket from the state
	TcpSocket *s = st->m_socket;
	// get it
	XmlDoc *xd = &st->m_xd;

	if ( ! xd->m_loaded ) {
		// setting just the docid. niceness is 0.
		//xd->set3 ( st->m_docId , st->m_coll , 0 );
		// callback
		xd->setCallback ( state , processLoop );
		// . and tell it to load from the old title rec
		// . this sets xd->m_oldTitleRec/m_oldTitleRecSize
		// . this sets xd->ptr_* and all other member vars from
		//   the old title rec if found in titledb.
		if ( ! xd->loadFromOldTitleRec ( ) ) return false;
	}

	if ( g_errno ) return sendErrorReply ( st , g_errno );
	// now force it to load old title rec
	//char **tr = xd->getTitleRec();
	SafeBuf *tr = xd->getTitleRecBuf();
	// blocked? return false if so. it will call processLoop() when it rets
	if ( tr == (void *)-1 ) return false;
	// we did not block. check for error? this will free "st" too.
	if ( ! tr ) return sendErrorReply ( st , g_errno );
	// if title rec was empty, that is a problem
	if ( xd->m_titleRecBuf.length() == 0 ) 
		return sendErrorReply ( st , ENOTFOUND);

	// set callback
	char *na = xd->getIsNoArchive();
	// wait if blocked
	if ( na == (void *)-1 ) return false;
	// error?
	if ( ! na ) return sendErrorReply ( st , g_errno );
	// forbidden? allow turkeys through though...
	if ( ! st->m_isAdmin && *na )
		return sendErrorReply ( st , ENOCACHE );

	SafeBuf *sb = &st->m_sb;


	// &page=4 will print rainbow sections
	if ( ! st->m_printed && st->m_r.getLong("page",0) ) {
		// do not repeat this call
		st->m_printed = true;
		// this will call us again since we called
		// xd->setCallback() above to us
		if ( ! xd->printDocForProCog ( sb , &st->m_r ) )
			return false;
	}

	char *contentType = "text/html";
	char format = st->m_format;
	if ( format == FORMAT_XML ) contentType = "text/xml";
	if ( format == FORMAT_JSON ) contentType = "application/json";

	// if we printed a special page (like rainbow sections) then return now
	if ( st->m_printed ) {
		bool status = g_httpServer.sendDynamicPage (s,
							    //buf,bufLen,
							    sb->getBufStart(),
							    sb->getLength(),
							    -1,false,
							    //"text/html",
							    contentType,
							    -1, NULL, "utf8" );
		// nuke state2
		mdelete ( st , sizeof(State2) , "PageGet1" );
		delete (st);
		return status;
	}

	/*
	  // this was calling XmlDoc and setting sections, etc. to
	  // get the SpiderReply junk... no no no
	// is it banned or filtered? this ignores the TagRec in the titleRec
	// and uses msg8a to get it fresh instead
	char *vi = xd->getIsFiltered();//Visible( );
	// wait if blocked
	if ( vi == (void *)-1 ) return false;
	// error?
	if ( ! vi ) return sendErrorReply ( st , g_errno );
	// banned?
	if ( ! st->m_isAdmin && ! *vi ) return sendErrorReply (st,EDOCBANNED);
	*/

	// get the utf8 content
	char **utf8 = xd->getUtf8Content();
	//long   len  = xd->size_utf8Content - 1;
	// wait if blocked???
	if ( utf8 == (void *)-1 ) return false;
	// strange
	if ( xd->size_utf8Content<=0) {
		log("pageget: utf8 content <= 0");
		return sendErrorReply(st,EBADENGINEER );
//.........这里部分代码省略.........