try
{
tuxConn = this.setUpConnection();
}
catch (TPException e)
{
System.out.println(\
throw e;
}
try
{
tuxRtn=tuxConn.tpacall(service_name,in_params, ApplicationToMonitorInterface.TPNOTIME);
}
catch (TPReplyException tre)
{
System.out.println(\
throw tre;
}
catch (TPException te)
{
System.out.println(\
throw te;
}
catch (Exception ee)
{
System.out.println(\
throw new TPException(TPException.TPESYSTEM, \
}
return tuxRtn;
}
函数Tpgetrply
public TypedFML32 Tpgetrply(CallDescriptor cd) throws TPException, TPReplyException
{
Reply tuxRtn = null;
TypedFML32 result = null;
try
{
if (tuxConn == null) tuxConn = this.setUpConnection();
}
catch (TPException e)
{
System.out.println(\
throw e;
}
try
{
tuxRtn = tuxConn.tpgetrply(cd, ApplicationToMonitorInterface.TPNOTIME);
}
catch (TPReplyException tre)
{
System.out.println(\
throw tre;
}
try
{
result = (TypedFML32)tuxRtn.getReplyBuffer();
}
catch (Exception e)
{
System.out.println(\
result = null;
}
tuxConn.tpterm();
return result;
}
关于避免调用超时
有时Tuxedo service需要做时间较长的处理,超过了tuxedo默认的调用超时
时间,为了避免超时需要使客户端一直阻塞知道服务器端service返回结果;在传统tuxedo的c语言程序客户端可以指定tp(a)call函数的第6个参数flag为TPNOTIME,此时客户端如果有阻塞条件存在,客户端会一直阻塞即使超时也不返回;但通过weblogic的java客户端如果想要达到这个目的,如果采用异步tpacall调用,不仅仅要在tpacall中指定flag值为ApplicationToMonitorInterface.TPNOTIME,在tpgetrply等待结果时也同样需要指定第2个参数为ApplicationToMonitorInterface.TPNOTIME,否则仅在tpacall中指定不能达到避免超时的目的.
编写ejb客户端