Infinite loop in WriteFile for a Serial Port

[g4j31a5]

Hi, guys. Long time no see. It's been like 1 year or so when I last posted here. Well, mainly in the C++ forum anyway. BTW, now I've got an issue. It's actually still a C++ one, but IMHO it's Windows related. I wanted to send a string to a COM port using the usual WriteFile command. But I don't know why, it's starting to loop endlessly (IMHO). The code is:

	DWORD len;
	DWORD dataLen = (DWORD) strlen( p_data );

	std::cout<<"Data to be sent:"<<p_data<<std::endl;

        if( WriteFile( m_hComm, p_data, dataLen,
		&len, NULL ) == 0 )
        {

		MessageBox( NULL, "NO: Failed sending data.", "Debug", MB_OK );
                return false;
         }
	MessageBox( NULL, "YES: Data sent.", "Debug", MB_OK );

When I started the code, it won't even display either one of the message boxes. What's wrong with it? Thanks in advance.

[Salem]

Where's the loop ?
Maybe call GetLastError and return some additional useful information in your fail messages.

[g4j31a5]

Where's the loop ?
Maybe call GetLastError and return some additional useful information in your fail messages.

The loop is in the WriteFile itself. So, I guess the word loop is not that accurate. More like it hangs if it reaches the WriteFile part. Anyway, it looks like I managed to solve it myself though. After I set the COMMTIMEOUTS.WriteTotalTimeoutConstant to 2000, it can exit the hangs after 2 secs. Sorry for the trouble guys.

[matsp]

Yes, if you don't set a timeout, it won't exit WriteFile until the data has been sent. In this case, I expect that you have some sort of flow control problem - perhaps your cable isn't wired up right, or you have one end set for CTS/RTS flow control, and the other end isn't set the same - so the other end never says "Ok, you can go ahead and give me some data now", and thus your WriteFile does exactly what it has been told to do.

--
Mats

[g4j31a5]

Yes, if you don't set a timeout, it won't exit WriteFile until the data has been sent. In this case, I expect that you have some sort of flow control problem - perhaps your cable isn't wired up right, or you have one end set for CTS/RTS flow control, and the other end isn't set the same - so the other end never says "Ok, you can go ahead and give me some data now", and thus your WriteFile does exactly what it has been told to do.

--
Mats

Yeah, I understand now. Thanks for pointing that out. BTW, a little OOT. I don't know what's wrong with the serial device but it seems like the input / output from / to the device is not always in the correct order. The weird thing is, I'm pretty sure that I've set the configuration just as the middleware programmer has told me (ie. 9600, 8N1, no flow control). Any other idea as of what may cause this issue? Thanks again.

[g4j31a5]

Nevermind. What seemed like an incorrect order is actually, IMHO, the difference of unicode mode between the application and the serial. I'll have to work it out myself I guess. Thansk anyway.