|
Hey, I've createt an axtiveX Control which shows a modeless dialog. Now my Control should notify the containerapp when there a changes in dialogs items. I tried to send a message from my dialogclass to the parent (ControlClass) but nothing happend. How can I notify my ControlClass from within my DialogClass?
|
|
|
|
|
You can notify the container application of an ActiveX control by fireing events to it. In order to create them, you should create an Event Map (very much like a Message Map in a "normal" MFC app) between your events and functions.
Look up BEGIN_EVENT_MAP, END_EVENT_MAP macros , and Add Event Wizard.
Peter Molnar
|
|
|
|
|
Thank you,
but only the Controll_Class can fire event's. My problem is to notify the controll_class from within my dialog_class. I tried to send messages but it didn't work. I created a PreTranslation -function in my controll-class, but no messages from the dialog_class reached the controll_class.
|
|
|
|
|
Sorry, I must have misunderstood you.
Your problem can then be led to communicating between modeless dialog and its creator thread.
For this you will have to be able to call the creator thread's functions, which is only possible if you pass a pointer of your creator thread to the modeless dialog. Using this pointer you post messages from the modeless dialog which can be caught in the control class.
Try the following:
1.Add to your modeless dialog (in the header file):
#define WM_SEND_MY_MESSAGE (WM_USER + 1)
COleControl* m_pCreator;
void inline SetCreatorThreadPointer(COleControl* pCreator) {m_pCreator = pCreator;};
2.Add to your control class a message handler:
ON_MESSAGE(WM_SEND_MY_MESSAGE, OnModelessDialogSend)
and also implement the function with prototype OnModelessDialogSend(WPARAM wParam, LPARAM lParam); From here you will fire your events.
3.In your control class:
m_pDialog->SetCreatorThreadPointer(this);
m_pDialog->Create(...);
4.In your modeless dialog, when your message occurs, post a message to your control class:
m_pCreator->PostMessage( WM_SEND_MY_MESSAGE, (WPARAM)"Hi!");
Peter Molnar
|
|
|
|
|
Hi guys
There is a simple way to open a new email with an attachement an the adress destination on it from an app when a button is pressed?
What about to have some text on it?
Best Regards
Doc
|
|
|
|
|
Have you looked at the MapiMessage structure? It is used by MAPISendMail() .
Five birds are sitting on a fence.
Three of them decide to fly off.
How many are left?
|
|
|
|
|
Without attachments but with To,CC, subject and body:
ShellExecute(NULL,"open","mailto:doctorpi@codeproject.com?cc=doctorpi@home.com&subject=Information&body=about opening default mail client",NULL,NULL,SW_SHOWNORMAL);
Peter Molnar
|
|
|
|
|
You do have the grasp on this; only you need to re-read your own message again: when you push something into a vector, a copy of whatever you push gets created. In your case, that is a copy of the pointer, not a copy of the object pointed to.
The offending delete call destroys the object, causing subsequent attempts to use the (now) stale pointer from the vector to go haywire.
Solution: Do not store the pointer to the object but the object itself; or store the pointer and make sure to delete as and when appropriate. That is, delete exactly once after the last use of the object in question.
Good luck,
Bernd
|
|
|
|
|
Right, you "copy" the pointer, but you don't copy what the pointer points to. If you call delete on pImageDisplay, then it deletes the CImageDisplay object, thus causing pImageDisplayVector[i] to point to unallocated memory. In other words, don't call delete until you're done with pImageDisplayVector. Just let pImageDisplay go out of scope. Then, in the destructor of whatever object pImageDisplayVector is part of, iterate through pImageDisplayVector calling delete on each item. This will destroy everything that's suppose to be destroyed.
|
|
|
|
|
Your responses really explained the situation! I was quite confused, thinking I had made a copy of the image itself, not just the pointer.........so actually deleting the vectorElement[i] will take care of freeing the memory on the heap since both were pointing to the same thing.. Many thanks for the clarification,
ns
|
|
|
|
|
Hello everbody. I have a strange problem. When I compile my dll (with Visual Studio .NET 2003) in Release the compiler give me those strange errors in linking:
error LNK2005: ___CppXcptFilter already defined in msvcrt.lib(MSVCR71.dll)
error LNK2005: ___xc_a already defined in msvcrt.lib(cinitexe.obj)
error LNK2005: ___xc_z already defined in msvcrt.lib(cinitexe.obj)
error LNK2005: ___xi_a already defined in msvcrt.lib(cinitexe.obj)
error LNK2005: ___xi_z already defined in msvcrt.lib(cinitexe.obj)
error LNK2005: __amsg_exit already defined in msvcrt.lib(MSVCR71.dll)
error LNK2005: _atoi already defined in msvcrt.lib(MSVCR71.dll)
error LNK2019: unresolved external symbol _main referenced in function _mainCRTStartup
The project property is the same that in debug. I don't know what is the problem. The fact is also more strange because if I compile in Release with another PC the compilation will succeed. The version of the Visual .NET is the same, but I don't know what are the differences.
Can you help me ?
Thank you very much for any answer.
|
|
|
|
|
Interesting.
Recheck project properties and make sure the DLL is set for multithreading and for dynamic linking.
Kuphryn
|
|
|
|
|
In order to link your project with any library I would reccomend you to use
#pragma commment(lib,"yourlib");
instead of adding libraries over Project/Properties/Linker/Additional dependencies, because #pragma's apply to both release and debug build settings, and therefore there is less chance that you forget to link with libraries in either build.
Peter Molnar
|
|
|
|
|
Thank you very much for the answer. I find the problem. I have a library linked to my dll and, in Release, it was too old. After a rebuild, all work right.
Thanks
|
|
|
|
|
Hello all,
I want to notify thread1 to end by itself. I don't like to use the function TerminateThread() to do this job. But thread1 is a worker thread not a UI thread so it doesn't have a message queue. I can't send some messages to it.
How can I do?
Thanks a lot
|
|
|
|
|
Well, if the worker thread waits for an event with WaitForSingleObject then
add one more event for quit event and then set that event by using SetEvent.
The thread should wait for signal from one of the both events by using
WaitForMultipleObjects API.
|
|
|
|
|
Thanks Dudi, I get an idea now.
|
|
|
|
|
Further to a recent discussion thread, I can now stream RGB type formats across a propriatory link and render the image remotely. This proves the concept, however RGB encoding is large so have shifted to H263 encoding and decoding. But of course it isn't generating a picture at the other end :-/ !!
The other end uses Graphedit with the appropriate decoder in place.
When I send the IMediaSample from the sending end I used to just send the sample length and the sample itself, I now also send the Synch , Preroll, discontinuity and Start Stop Reference times to ensure the MediaSample I generate at the other end is representative, ie prior to going into the H263 decoder... but no joy.
Has anyone attempted the same thing and had problems with H263 (have also tried H261) or got any ideas?
Many Thanks
Loz
|
|
|
|
|
hi all,
am stuck with a crucial and interstng problem.
if anyone can help in solving TSP(Travelling sales person) problem using the Branch and Bound Algorithm.?
thanks
Deepesh Dhakad
Indian Institute of Information Technology,Allahabad
India
|
|
|
|
|
Have you seen:
http://www.math.princeton.edu/tsp/
http://www.cs.rutgers.edu/~chvatal/tsp.html
http://www.google.com/search?hl=en&ie=UTF-8&oe=UTF-8&q=traveling+salesman&btnG=Google+Search
http://www.nist.gov/dads/HTML/travelingSalesman.html
http://www.cis.ohio-state.edu/~gurari/course/cis680/cis680Ch20.html
http://www.idsia.ch/idsiareport/IDSIA-09-02.pdf
Five birds are sitting on a fence.
Three of them decide to fly off.
How many are left?
|
|
|
|
|
Hello folks.
I'm currently porting TCP/IP server program from UNIX framework. It uses classical scheme "one connection - one child process". Here's the code snippet:
retval=select(backend+1,&rfds,NULL,NULL,&tv);
if(retval>0) {
if(FD_ISSET(backend,&rfds)) {
if(peer=accept(blablabla)==0) {
if((pid=fork())!=-1) {
/* this is the child */
/* and he processes
* the request */
}
/* father skips the above section and fetches the next connection */
}
How do I implement this in Windows ?
fork() is absent, only spawn() which is
unusable here. Doing everything in single
process (even threaded) is not an option for me. How's the classical Windows internet server works ?
Max.
|
|
|
|
|
You don't: in *nix, this is a common practice, but in Windows, this is horrible, because processes have a overhead much greater than threads. So, in Windows we don't fork(), we CreateThread().
IOW, you won't see lots of httpd running on a Windows web server machine, you'll see a single inetinfo.exe.
Actually, if you need to port a *nix application to Windows, I strongly suggest you to use the free Cygwin, which, among other things, will give you fork().
On MSDN website there's an article series somewhere with guidelines and tips on porting *nix programs to Windows. I strongly suggest you to read it.
Trying to make bits uncopyable is like trying to make water not wet.
-- Bruce Schneier
By the way, dog_spawn isn't a nickname - it is my name with an underscore instead of a space. -- dog_spawn
|
|
|
|
|
The very highest performance Windows server applications don't create a thread for each request, they use a pool of pre-created threads which wait for requests to arrive. This can be done using the completion port API.
I suggest consulting Jeffrey Richter's book Programming Server-Side Applications for Windows 2000.
|
|
|
|
|
Mike Dimmick wrote:
The very highest performance Windows server applications don't create a thread for each request, they use a pool of pre-created threads which wait for requests to arrive
In the same way highest performance *nix servers don't fork() a CGI on every request and keep a pool of pre-forked processes.
Trying to make bits uncopyable is like trying to make water not wet.
-- Bruce Schneier
By the way, dog_spawn isn't a nickname - it is my name with an underscore instead of a space. -- dog_spawn
|
|
|
|
|
One other and probably worse way than writing as a windows process is to use the POSIX library MS has for Windows NT Systems which allows you to use your existing Unix code to create and run applications.
|
|
|
|