C / C++ / MFC Discussion Boards

Re: using SysListView32 in a generic win32 dialog app and using MFC classes without MFC

Rassman2-Nov-01 2:42

2-Nov-01 2:42

did you show the window after creating it ShowWindow(SW_SHOW) and is your window class registered.

Some of the MFC classes can be used standalone, particularly the those that do not include a view. CTime and CString are typical ones that would be used.

I haven't done it with a Socket, but I have taken some diagnostic ones. It is a matter of including the right headers (and sub headers) and setting the right #defines to enable the parts your functions need (otherwise done for you by a wizard). You can see a list of these in .h files and the right Linked dlls. Remember that the controls are all based on CObject too so as far as 'visible' objects are concerned you are not able to ignore the MFC4xx.dll's altogether, I'm afraid I don't have my list of whats in each one anymore.

We do it for the joy of seeing the users struggle.

I have a property sheet in a CFormView with 5 property pages, but they are the same pages of content. Any way of reusing code?

DanYELL1-Nov-01 8:58

DanYELL

1-Nov-01 8:58

I have a CFormView and inside it is a CPropertySheet with
5 property pages. The property pages are identical. I typed:

CMyPropertySheet.h
------------------

#include "MyPage1.h"

...

CMyPage1 page1;
CMyPage1 page2;
CMyPage1 page3;
CMyPage1 page4;
CMyPage1 page5;

and then I added the pages. It works, but for awhile. Then I
get a Fatal Exception Handling error and the computer just goes
nuts! I have to reboot. I know I am doing something wrong, but
I don't know how to resolve it unless I have 5 separate classes
and 5 separate IDD_PAGE1, IDD_PAGE2, rather than just 1.

Any way of reusing code without getting massive errors?

Please, any response any one can give me will be greatly
appreciated.

Sincerely,
Danielle (an overworked graduate student)

TCP/IP support for DOS

1-Nov-01 8:51

I am running DOS on my embedded system and I need to be able to send data back through the net. I am using Borland Turbo C. Does anyone know of any code that I can use to handle the TCP\IP stack, like Windows Socket API for Win32?

Please help.;)

Re: TCP/IP support for DOS

Nish Nishant1-Nov-01 17:01

Nish Nishant

1-Nov-01 17:01

take a look at http://www.alumni.caltech.edu/~dank/trumpet/

Nish

Re: TCP/IP support for DOS

Rassman2-Nov-01 2:52

Rassman

2-Nov-01 2:52

The link from Nish sums it up well.

From DOS 5 upwards there is built in calls to help in this sort of communications. I have a DOS Bible (somewhere arround here) if you can get hold of one of those then you will see a list of all DOS calls and the parameters used. You can speedup your embedded systems no end if you use these rather than allowing the C interface to do it.

A typical one that the C compilers did was to use the indirect functions of DOS to write to the screen, it is legal in DOS to use one its more direct screen writeing calls.

There is also a FileIO function (off the top of my head call 24 or 42) that covers just about any communications you could need, including IP.

We do it for the joy of seeing the users struggle.

Named pipes over network from COM/IIS

bburton1-Nov-01 5:53

bburton

1-Nov-01 5:53

I have the need to access a named pipe over a network where both sides are Windows NT 4.0 machines. The pipe server is implemented as an NT service while the client request will originate from an ATL COM dll. I have zero problems if I invoke the COM object from my test application; the two machines "talk" to each other, transfer data back and forth, etc.

My problem is when the COM object is invoked as a result of IIS activity. In this case, my cgi application (written in VC++) creates the COM object. The COM object then attempts to connect to the pipe using CreateFile(). I am sure this is a problem with the security attributes setting for the client side, at this point the arguement is set to NULL.

BTW - the named pipe is created WITH security attributes... A SD is declared, then initialized using InitializeSecurityDescriptor() and set using SetSecurityDescriptorDacl(). sa.lpSecurityDescriptor is then set to the SD.

So, any ideas on how I might resolve this little problem??

TIA,

Barry

Re: Named pipes over network from COM/IIS

Martin Ziacek1-Nov-01 7:07

Martin Ziacek

1-Nov-01 7:07

Barry, I think security attributes defined by your server prohibit access to pipe. When you create named pipe (server side) you specify, who can access it. However, local system account from another machine has no right to access resources on the remote machine (your scenario, when COM object created by IIS access named pipe).

I would recommend two solution:

1. set security attributes to allow access for everyone (bad, not secure solution)
or
2. impersonate client COM object to proper user account to access remote resources (your named pipe).

I hope this will help you.

Kind regards.

Martin

Re: Named pipes over network from COM/IIS

Anders Molin1-Nov-01 9:58

Anders Molin

1-Nov-01 9:58

When you "execute" a COM object from a cgi application, it is running in the local anonymous web-user's security context (IUSER_something), that user normally dont have the rights to access the remote machine. The problem is the same with the system account when running from a server.

The most easy way, AFAIK, is to get people to log on the webserver, instead of being anonymous users, then the COM object executes in their security context, and then it might have the rights to access the remote server.

I *think* there is a way to say to IIS that "this application (your CGI app) should run as this user", but right now I don't remember how to do it...

You can also say that anonymous users on IIS should run as a domain user that you select, then they have the rights to access other computers on the network, but then you have a huge security risk...

- Anders

Money talks, but all mine ever says is "Goodbye!"

Bit operation question

1-Nov-01 5:20

Hi
Does anybody knows how to make the following only with bit operations

if( a1==b1 )
a1 = c1;

Re: Bit operation question

Martin Ziacek1-Nov-01 6:55

Martin Ziacek

1-Nov-01 6:55

Is this sufficient solution?

if (!(a1 ^ b1)) a1 = c1;

Re: Bit operation question

1-Nov-01 7:04

Thank you, but I do not want to use if

Re: Bit operation question

Martin Ziacek1-Nov-01 7:44

Martin Ziacek

1-Nov-01 7:44

What about:

a1 = !(a1 ^ b1) ? c1 : a1;

BTW, what's the reason to not use 'if'?

Re: Bit operation question

C^hris L^osinger1-Nov-01 7:44

C^hris L^osinger

1-Nov-01 7:44

can you use arithmetic operators?

http://www.smalleranimals.com

one with arithmetic operators:

C^hris L^osinger1-Nov-01 7:52

C^hris L^osinger

1-Nov-01 7:52

((a==b) * c) | ((a!=b) * a)

-c

http://www.smalleranimals.com

Re: one with arithmetic operators:

Alvaro Mendez1-Nov-01 7:57

Alvaro Mendez

1-Nov-01 7:57

Chris,

I don't see where the assignment of a to c1 takes place in your expression. Could you clarify, or is it just a joke?

Regards,
Alvaro

Re: one with arithmetic operators:

C^hris L^osinger1-Nov-01 7:58

C^hris L^osinger

1-Nov-01 7:58

oh sorry, that's: a = ((a==b) * c) | ((a!=b) * a);

http://www.smalleranimals.com

DIBS and "the bits"

Chambers1-Nov-01 5:11

Chambers

1-Nov-01 5:11

Hi all, I have managed to load/create (as necessary) a 24-bit DIBSection using the means described on this very web site (LR_CREATEDIBSECTION flag in LoadImage and CreateDIBSection), both of which return a HBITMAP. Superb. What I want to do now is "grab" the actual BGR colour values that make up each pixel, and change them to something else such as RGB(0,255,0 - Green, its the easiest example to use). At the moment I am trying to get the bits using a BITMAP structure (returned through GetObject) and its bmBits property. I have also implemented Christians wizardry to convert this pointer to a valid LPBYTE by :

for (int y=0, y < bm.bmHeight, y++)
{
  LPBYTE pbSrcRGB = (LPBYTE)&((DWORD*)pDestBits)[y*bm->bmWidth]
  for (int x=0, x < bm.bmWidth, x++)
  {
    pbSrcRGB[0]=0
    pbSrcRGB[1]=255
    pbSrcRGB[2]=0
    pbSrcRGB+=3
  }
}

However, the program crashes out after a few laps of the for loops. I have also checked to see whats in pbSrcRGB at position [0] BEFORE assignment and every time it has :

0 ''

it is NULL terminated. How is this remedied? BITMAP structure is passed into this function by :

&bm

and a LPBITMAP structure is then used (as you can see from bm->), so I would have thought that everything was OK from the BITMAP point of view. Also the bitmap starts off black (which I don`t want, thats why I`m changing it all to green), so I assume this is why it reads 0 ''. Also when I check pbSrcRGB[0] to see what it contains AFTER assignment, it reads :

255 'y' (with funny line over the top of the y)

Why does this happen?

The main essence of my questions are aimed at discovering :
a) why the program could be crashing out.
b) why the pixel colours on the screen don`t change when the memDC is blit to pDC (the screen device context).

Any help would be hugely appreciated as you can imagine. I`ve been working on this for a while, and just can`t figure it out. I`ve checked a lot of articles, but to this day I can`t find one that even attempts to shed light on the situation (except Christians - but the article didn`t really target the issue). If you`ve read this far down, I thank you for that even if you don`t know the answer,

Cheers, Confused | :confused:

Alan.

P.S. On a slightly different aspect, is there anyway of creating a DIB and intialising its colour to white instead of black?

AEGC

Re: DIBS and "the bits"

C^hris L^osinger1-Nov-01 5:21

C^hris L^osinger

1-Nov-01 5:21

try using the <pre>...<pre> tags around code sections. and use < for &"<" characters.

about the NULL-terminated and funny-Y characters - don't worry about them. the debugger is displaying the BYTEs as characters.

http://www.smalleranimals.com

Re: DIBS and "the bits"

Chambers1-Nov-01 5:36

Chambers

1-Nov-01 5:36

Cheers for page authoring advice, it looks much prettier now. So you think everything seems fine, but why does the app crash out? and why do the colours not change?

Cheers for the response,
Alan.

AEGC

Re: DIBS and "the bits"

C^hris L^osinger1-Nov-01 5:57

C^hris L^osinger

1-Nov-01 5:57

this code might help:

// find row padding
int nPad = ds.dsBm.bmWidthBytes - (((ds.dsBmih.biWidth *
        ds.dsBmih.biBitCount) + 7) / 8);

LPBYTE pbBits = (BYTE*) ds.dsBm.bmBits;
for (i=0; i<ds.dsBmih.biHeight; i++) 
{
   for (j=0; j<ds.dsBmih.biWidth; j++)   
   {
      *(pbBits+0) = 0;
      *(pbBits+1) = 255;
      *(pbBits+2) = 0;
      pbBits+=3;
   }
   pbBits += nPad;
}

http://www.smalleranimals.com

Re: DIBS and "the bits" - Chris Please Read

Chambers1-Nov-01 6:12

Chambers

1-Nov-01 6:12

Well, you truly are a magician of the C++ variety, the code example you gave me worked superbly. However, there is one slight problem which is preventing me from tarring you with the same golden brush as Gandalf the Great. The bitmap goes green alright, but for only a second, and then where I draw some lines in some code after it, it flicks back to bl**dy black. Please enlighten me.

Also, I just like to thank you for that piece of code you gave me, for the page authoring tips you mentioned, and, well...for bothering to reply at all.

Cheers mate, I owe you one,
Alan. Smile | :)

Re: DIBS and "the bits" - Chris Please Read

C^hris L^osinger1-Nov-01 6:58

C^hris L^osinger

1-Nov-01 6:58

hmmm... not sure about this one. are you re-drawing the bitmap anywhere?

-c

http://www.smalleranimals.com

Re: DIBS and "the bits" - Solution

Chambers1-Nov-01 7:33

Chambers

1-Nov-01 7:33

I have finally found what was going wrong! I had selected the wrong hbitmap into the DIBSECTION (it was the background layer), thats why it was flicking up so quickly, it was being replaced by the new black background I just created. Just thought I`d post the solution in case anybody else has a similar problem. My thanks go Chris on majorly helping me through the subject. I`ll also thank Christian for his part on getting me on the right road to success.

Cheers guys,
Alan. Smile | :)

Accelerator Keys

Özgür1-Nov-01 4:51

Özgür

1-Nov-01 4:51

Hi,
I have a MDI Application with threeway splitter ChildFrames.
I inserted some shortkeys through resource editor.
Shortkeys are invoked only when the pane holding the control(related to shortkey) has the focus.
I want shortkeys to be invoked without caring about the focus..
How can i do it?
Can anybody put some light infront of me?

with greetings from germany
özgür

sorting std::vector

User 98851-Nov-01 4:23

User 9885

1-Nov-01 4:23

I have a class that has a member variable of type vector<CMyClass >.
I have sort parameters stored in the class that I want to use inside a custom sort function. But, the sort function has to be either global or a static member function. How can I access the class member that defines the sort criteria from the sort function. I cannot pass it as a variable also. The only solution that I could think of was to use a seperate function for each sort type, but this results in too many functions. I have 28 sort criteria in this class.

Thanks in advance for any help

-Thomas

modified 29-Aug-18 21:01pm.

Use Ctrl+Left/Right to switch messages, Ctrl+Up/Down to switch threads, Ctrl+Shift+Left/Right to switch pages.