[RESOLVED] Why doesn't this method of closing a worker thread work?

Quote:

---

Originally Posted by 2kaud

The easiest way is probably to sleep the thread for 50ms.

---

This only guarantees a lower limit. So there is a possibility that the process might sleep for longer and miss the next packet of data. Won't my idea work?

If the data is only being received every 50ms, then its probably easier to just wait for the data comm event to occur using WaitCommEvent(). See https://msdn.microsoft.com/en-us/lib...(v=vs.85).aspx

This will wait until a comm event is signalled. The lpEvtMask parameter indicates what comm event is signalled. For data being received then this
would be EV_RXCHAR when all chars available could then be read. The loop then turns into something like

Code:

---

DWORD        dwCommEvent;

while (true) {
    WaitCommEvent(hcomm, &dwCommEvent, NULL);
    if (dwCommEvent & EV_RXCHAR) {
        //Character(s) waiting to be read
        //Read all available chars from comm port and process
    }
}

---

A code example of reading from a comm port, assembling one line of read data and processing it is

Code:

---

//Thread master function
unsigned __stdcall ProcessData(PVOID args)
{
HANDLE        hComm = *(HANDLE*)args;

DWORD        dwCommEvent,
                dwRead;

char        chRead;

string        line;

        for ( ; ; ) {
                if (WaitCommEvent(hComm, &dwCommEvent, NULL) && (dwCommEvent & EV_RXCHAR))
                        do {
                                if (ReadFile(hComm, &chRead, 1, &dwRead, NULL)) {
                                        if (dwRead == 1) {
                                                if (chRead < ' ') {
                                                        if (chRead == 13) {
                                                                Process(hComm, line);
                                                                line = "";
                                                        }
                                                } else
                                                        line += chRead;
                                        }
                                } else {
                                        cout << "error in the ReadFile call\n";
                                        return 8;
                                }
                        } while (dwRead);
                else {
                        cout << "error in waitcommevent\n";
                        return 9;
                }
        }
}

---

Thank you so much! This is way beyond anything you needed to do!
Could you explain what you are doing with chRead in the two if statements? Also what does Process do? I just want to understand this, rather than just lift it.
A link to any resource would do, I would not want to waste your time.

Another good essay of Joe Newcomer about using serial ports: http://flounder.com/serial.htm

Quote:

---

Could you explain what you are doing with chRead in the two if statements?

---

hComm is the handle to the port obtained from CreateFile().

When a comms event occurs then WaitCommEvent() will return. If the event is EV_RXCHAR then at least one char is ready to be read from the port. The do loop reads one char at a time from the port until no chars are read. For the purpose this code is used, the device being read sends data one line at a time terminated by a CR (ASCII code 13). If a char 13 is read then the end of line has been read and the accumulated string is passed to the process function. Chars with a value of less than a <space> in this application are ignored - hence the test for < ' '. If the read char is >= ' ' then it is appended to the string line and the loops repeat. Whether the processing of the read char in this example code in your case is appropriate or not will depend upon the format that the device from which you are reading sends its data.

Quote:

---

Also what does Process do

---

Process() simply processes the completed assembled line of data read from the device in the string line. This is whatever you want to do with the read data.

Cheers!

Quote:

---

Originally Posted by VBCherry

This only guarantees a lower limit. So there is a possibility that the process might sleep for longer and miss the next packet of data.

---

No Windows user timer can be relied upon to be 100% accurate as the OS is not real-time. I agree that having Sleep() in code is not usually a good idea but I thought that you wanted simply to sample the input every 50ms rather than read the data which is received every 50ms. IMO using any windows timing mechanism in this scenario is not recommended as exactly 50ms cannot be guaranteed - and if slightly longer could miss the next packet as you stated. Also the time taken to process the data packet needs to be taken into consideration (which needs to be less than 50ms). The wait time plus the process time needs to be 50ms which is going to be very difficult - if not impossible - to undertake. Therefore as suggested in my posts #17 and #18 and in the article referenced by Victor in post #20 IMO other ways should be investigated that don't rely upon timing.