🚀 go-pugleaf

Hi Gambit,

Thanks for trying to help.


Remy Lebeau (TeamB) wrote:
> "Tassilo J. Klein" <nospamTJKlein@web.de> wrote in message
> news:dctdlr$6gr$02$1@news.t-online.com...
> 
> 
>>Invocation of CloseHandle function out of any reason then calls the
>>destructor of the calling TThread ~DownloadThread()
> 
> 
> The only way that could be happening is if CloseHandle() is throwing an
> exception that is terminating the thread, and if you have the thread set to
> FreeOnTerminate=true (or external code that is 'delete'ing the thread after
> a call to WaitFor()).
> 
> 

Yes, is FreeOnTerminate = true;


>>although CloseHandle is encapsulated by try/catch blocks.
> 
> 
> That does not necessarily mean that the try..catch will actually catch an
> exception.  It depends on what your catch block is actually looking for, and
> what kind of exception, if any, is actualy being thrown.  What does the
> actual code look like?

Here is a part of the code of the TThread. But everything critical is 
encapsulated with try and catch(...). I don't understand it....

void __fastcall DownloadThread::Execute()
{

     while( Terminated == false)
     {

         try
         {	// opening the port to the device

		if ( device->startDevice() != DEVICE_OK )
                         throw(TRANSFER_ERROR);

		// download data from the device

		if ( device->downloadSpectrum() != DEVICE_OK)
                         throw(TRANSFER_ERROR);
		
		.....
	}
	catch(...)
	{
		...
		// close the port
		device->stopDevice();
		...
	}
	WaitForSingleObject(event, wait );
}

// stop the device, close the port
virtual void TMDU::stopDevice()
     {
         closePort();
         monitor->Release();
     };


void TMDU::closePort()
{
     //closing the COM port
     try
     {
     // write a byte to the device to finish communication
     exitPort();

     flushPort();

      if ( hCom > 0 )
     {
      // clear data in the puffer
       PurgeComm(hCom, PURGE_TXABORT | PURGE_RXABORT );

       // free the handle
      CloseHandle( hCom ); }              <-------- from here the 
debugger steps to the destructor of the download thread

     hCom = NULL;
     }
     }
     catch(...)
     {
      .....
     }
}


Any clue?

Regards,
  -Tassilo

> 
> 
> Gambit
> 
>