C / C++ / MFC

Re: How to instantiate class with parameters - syntax please ?

4-Oct-16 7:16

Sorry my mind reading was a bit off, I am however glad you worked out you were trying to build a composite class.

The forum can help with specific small problems but it's not something that can teach complex techniques in C++.

In vino veritas

Vaclav_6-Oct-16 10:34

Vaclav_

6-Oct-16 10:34

Yes, this time using correct terminology was the ticket. I am not sure why the OP was judged as too complex to ask here about. I am still puzzled why I did not get a clear example of syntax passing list of parameters to class directly, not via constructor. The link posted was not that clear how to name / use the list.
It really does matter to me how I get the parameters to correct class in the hierarchy.
The example of composite class suggest I have to follow the hierarchy and pass the parameters thru each constructor to the next one.
I suppose when I get the code to play I could post it here for everybody can see it is really not that complex.
Cheers

Re: How to instantiate class with parameters - syntax please ?

leon de boer6-Oct-16 19:04

Why can't I assign a variable's address to an enum?

6-Oct-16 19:04

Correct

The problem as you will find is hierarchy order as you will find. Sometimes the order will not
be the way you want in different functions and you will need to control it and that gets down
to some very nitty details.

The question we will still always pose back to you is do you really need the hierarchy at all.

I can't stress enough to read one article from wikipedia it can save you alot of heartache
https://en.wikipedia.org/wiki/Composition_over_inheritance[^]

In vino veritas

arnold_w30-Sep-16 0:18

arnold_w

30-Sep-16 0:18

I am writing standard C programming language for an ARM processor and I would like to combine a handle (a pointer to a struct with state variables) with an enum like this:

C++

struct MyStateStruct_t {
    int a;
};

const struct MyStateStruct_t myStates [2];

typedef enum
{
    MY_SELECTOR_0  = (const int)(&myStates[0]),
    MY_SELECTOR_1  = (const int)(&myStates[1])
} MyEnum_e;

When I try to compile, I get the following errors:
enumerator value for 'MY_SELECTOR_0' is not an integer constant
enumerator value for 'MY_SELECTOR_1' is not an integer constant

Does anybody know a workaround? Of course, I could use a look-up table array to convert from the enum-values to the addresses, but if possible I would like to avoid that. Thanks for any help!

Re: Why can't I assign a variable's address to an enum?

Richard MacCutchan30-Sep-16 0:52

Richard MacCutchan

30-Sep-16 0:52

Enums must be defined integer constants at compile time.

Re: Why can't I assign a variable's address to an enum?

leon de boer30-Sep-16 3:22

WaitForSingleObject best practices

30-Sep-16 3:22

As Richard said enum is

1.) Restricted to a 16 bit integer constant
2.) Must be defined at compile time

Your pointer is longer than 16 bits and the address is only available at runtime, so it double fails.

You also haven't typedef the structure to a name.

I suspect this is what you are trying to do

typedef struct MyStateStruct_t {
    int a;
} MYSTATE;        // typedef the structure to a name

typedef enum {
    MY_SELECTOR_0 = 0,
    MY_SELECTOR_1 = 1
} MyEnum_e;

const MYSTATE myStates[2];   // use structure name

enum MyEnum_e selector = MY_SELECTOR_0;  // Set selector to 0
int result1 = myStates[selector].a;       // result1 is myStates[0].a

selector = MY_SELECTOR_1;  // Set selector to 1
int result2 = myStates[selector].a;       // result2 is myStates[1].a

In vino veritas

ForNow27-Sep-16 14:26

27-Sep-16 14:26

Hi

I am looking to be Notified When something is downloaded to my FTP folder
towards that end I have implemented FindFirstChangeNotification being a mainframe programmer its taken a while to get used that when I do a Wait I stop the entire thread

So I have 2 options put the wait in a different thread or.... code a timer value other then infinite and in the case of WAIT_TIMEOUT loop "while(TRUE)" and Only "break when
the object is signaled

leon de boer27-Sep-16 18:18

27-Sep-16 18:18

Usually you don't stop an entire thread while waiting you are forced to do it with MFC, so I am guessing this is your MFC program again.

Within MFC personally I would create a thread which just spins it wheels checking whatever and the thread simply posts a message back when an interesting event occurs. I wouldn't be putting the wait I would be putting the whole code that requires the wait in a thread and use messages to communicate with the MFC code. So make a proper worker thread.

In vino veritas

ForNow28-Sep-16 2:18

28-Sep-16 2:18

So you are saying creating a new thread to wait for notification is a better choice then an ifinate loop with say
A 1000 millisecond in the WaitForSingleObject only breaking out of the loop when the event is signaled
Truth is a new thread leaves the CPU time distribution up to Windows

Thanks

leon de boer28-Sep-16 8:16

28-Sep-16 8:16

It's a low priority Hotwatch .. it will always be better to do as a thread and it isn't hard

// Define a structure to pass to thread
typedef struct _MYWATCHERDATA {
	HWND AppWindow;             // Application window to post message to
	UINT MsgId;			        // Your private message ID will be WM_APP + some number
	WPARAM CmdId;               // Command ID .. perhaps you have multiple of these watchers
	char* FolderToWatch;        // The string nominating the folder to watch
	HANDLE dwChangeHandles[2];  // These will be thread wait handles (1 for it, 1 for me to force exit)
	BOOL threadExitComplete;	// Thread exit completed
} WATCHERDATA;

// This is the thread routine
// Pass an initialized structure in on LPARAM Of CreateThread
DWORD WINAPI NotifyMeOfFolderChangeThread(LPVOID lpParam)    // thread procedure
{
	BOOL exitSignal = FALSE;                                 // Preset exit signal to false
	WATCHERDATA* watchData = (WATCHERDATA*)lpParam;	         // Typecast the lParam to your structure you pass in

	// Watch Handle[0] belongs to this thread 
	watchData->dwChangeHandles[0] =
		FindFirstChangeNotification(watchData->FolderToWatch, // folder path passed in
			FALSE,										      // no subfolders
			FILE_NOTIFY_CHANGE_FILE_NAME);				      // watch for renaming, creating, or deleting a file
	
	if (INVALID_HANDLE_VALUE == watchData->dwChangeHandles[0])// Error something is dorked probably folder doesn't exist
	{
		DWORD dwError = ::GetLastError();
		// handle error
		return dwError;
	}

	while (exitSignal == FALSE) {
		// Okay we wait on the notification signal or exernal exit signal
		DWORD dwWaitStatus = WaitForMultipleObjects(2, 
			watchData->dwChangeHandles, FALSE, INFINITE);
		
		switch (dwWaitStatus) {
			// Wait exit was from directory change
			case WAIT_OBJECT_0:
				// Post off message back to application window.
				PostMessage(watchData->AppWindow, watchData->MsgId,
					watchData->CmdId, (LPARAM)watchData->FolderToWatch);
				FindNextChangeNotification(watchData->dwChangeHandles[0]);
				break;
			// External thread signal forcing thread to break out and exit
			case WAIT_OBJECT_0 + 1:
				exitSignal = TRUE;           // Yep time to exit
				break;
		}
	}
	watchData->threadExitComplete = TRUE; // Thread has finished with watchData .. tell app that
	return 0;
}

The use of the thread involves, initialize, use, wait for thread to exit .. cleanup

// Initialize a WATCHERDATA structure
char* watchDirectory = "c:\\temp\\";                                // Some directory to watch
WATCHERDATA* myWatch = (WATCHERDATA*) malloc(sizeof(WATCHERDATA)); // Allocate memory
myWatch->AppWindow = myAppHandle;                                  // Application window handle
myWatch->MsgId = WM_APP  + 100;									   // Some private application message ID 
myWatch->CmdId = 1;										           // Some private id we might have multiple watches
int len = strlen(watchDirectory) + 1;                               // size of string
myWatch->FolderToWatch = (char*) malloc(len);					   // Allocate string space 
strcpy_s(myWatch->FolderToWatch, len, watchDirectory);			   // Copy string to allocated memory        
myWatch->dwChangeHandles[0] = 0;								   // Clear thread handle one .. thread itself uses that
myWatch->dwChangeHandles[1] = CreateEvent(0, FALSE, FALSE, "EXIT");// Set thread handle two for us to force exit
myWatch->threadExitComplete = FALSE;	                           // Thread will set this on complete
																   
// Use thread
HANDLE workerThread = CreateThread(NULL, 0, NotifyMeOfFolderChangeThread, myWatch, 0, NULL);

// Change thread priority
SetThreadPriority(workerThread, THREAD_PRIORITY_LOWEST);

// This will do a simple 10 sec wait .. simulating doing something
Sleep(10000);

// Ok manually terminate the thread
SetEvent(myWatch->dwChangeHandles[1]);

// Wait for thread to signal it has finished with myWatch
while (myWatch->threadExitComplete == FALSE) {};

// Okay you are free to cleanup myWatch
free(myWatch->FolderToWatch);
free(myWatch);

In vino veritas

ForNow28-Sep-16 11:41

28-Sep-16 11:41

thanks so much

ForNow28-Sep-16 15:57

28-Sep-16 15:57

Leon

Just wanted to tell you again you are correct I have Sockets in my code

So Typically after I do CAsynSocket connect I do the send after I get the OnSend Notification

From The FrameWork

My Sockets has a CWinThread Wrapper, However the notification like OnSend comes in

The Context of the Main Thread

Since I put WaitFor...Object api (In The Main Thread) even though I didn't have INFINITE

it messed up the Windows Dispatcher and I wasn't getting the notification

I am going to try your code out Shortly

Thanks again

leon de boer28-Sep-16 19:20

28-Sep-16 19:20

Yes that is a generic schema for not doing that sort of thing ... I hacked the MSDN thread sample.

When I do socket apps I do exactly the same schema I have a write thread and a read thread of exactly
the form above. The socket IO structures and objects are held in the memory block.

Basically you want your app just dealing with messages not sitting around doing things .. bring up task
manager and look at the number of threads running normally. It really expects and wants this sort of
behaviour, it's designed around it.

Lifetime of app threads in particular are extremely easy because you can launch them on the WM_CREATE
of the main app window and dispose of them in the WM_DESTROY. You don't have the problem of who is
responsible for the shared data and memory (and it's cleanup) that you have with intermittent start
and stop threads. The app window handle to pass to the thread to message you also can't get wrong Smile | :)

They give you huge blocks of message ids at WM_APP and WM_USER to allow you to do it. All those ID's
you are free to use but you might want to check MFC might use a few (that is why I jumped 100) and
they may restrict the range down a little.

In vino veritas

leon de boer30-Sep-16 9:42

30-Sep-16 9:42

Present for you as I checked my code putting it into an application

Code: Folder watcher code[^]
Compiled EXE:Folder HotWatcher EXE[^]

It's a full folder watcher app with a basic watcher thread and detailed watcher thread.
Basic just tells you if the folder changed, Detailed tells you file changes and what action was done.

Basically run the app select an unimportant directory like c:\temp and then copy, delete, rename files to the directory and you will see the messages reported to the app.

In vino veritas

ForNow30-Sep-16 9:52

30-Sep-16 9:52

leon thanks so much I implemented the code and problems went away meaning with out the wait in the Main Thread The Framework was able to notify me via CAsynSocket::OnSend

The actual testing will come when I execute my Z/OS (MainFrame) FTP code to download the file

to that folder that is being watched

Much thanks

Joe Woodbury24-Oct-16 14:31

Joe Woodbury

24-Oct-16 14:31

Be aware that the change folder notification occurs at any change, which means the full write may not have happened. One solution is to write the file to another location and then atomically move it to the monitored directory.