Re: firewall by-passing - C / C++ / MFC Discussion Boards

Re: CString::GetLength(), what does it in fact return?

Graham Bradshaw15-Aug-05 9:08

Graham Bradshaw

15-Aug-05 9:08

Are you sure about this? I just tried it with VC6, and in a Unicode build, the code

CString strText = _T("abcde");<br />
<br />
long lLength = strText.GetLength();<br />
<br />
ASSERT(sizeof(TCHAR) == 2);<br />
ASSERT(lLength == 5);

does not ASSERT.

Re: CString::GetLength(), what does it in fact return?

David Crow15-Aug-05 9:27

David Crow

15-Aug-05 9:27

Graham Bradshaw wrote:
Are you sure about this?

No. I misread the help for GetLength(); when it was talking about MBCS, I was seeing Unicode!

I guess if I had looked at some of my code first, this would have clued me in that it is the number of characters:

...str.GetLength() * sizeof(TCHAR);

"One must learn from the bite of the fire to leave it alone." - Native American Proverb

Re: CString::GetLength(), what does it in fact return?

Jose Lamas Rios15-Aug-05 9:11

Jose Lamas Rios

15-Aug-05 9:11

DavidCrow wrote:
It returns the number of bytes in the object. For Unicode, this is twice the number of characters.

Actually, for Unicode, the returned length is the count of WCHARs. In other words, it's always the length in TCHARs (bytes for non-Unicode, WCHARs for Unicode). So, CString(_T("Four")).GetLength() will always return 4, regardless of Unicode settings.

The semantics for CString::GetLength() are the same as for the lstrlen[^] function, which is used internally by CString implementation. For instance, see the implementation for CString::ReleaseBuffer() (VC 6.0)

void CString::ReleaseBuffer(int nNewLength)
{
   CopyBeforeWrite();  // just in case GetBuffer was not called

   if (nNewLength == -1)
      nNewLength = lstrlen(m_pchData); // zero terminated

   ASSERT(nNewLength <= GetData()->nAllocLength);
   GetData()->nDataLength = nNewLength;
   m_pchData[nNewLength] = '\0';
}

--
jlr
http://jlamas.blogspot.com/[^]

Re: CString::GetLength(), what does it in fact return?

Tim Smith15-Aug-05 13:30

Tim Smith

15-Aug-05 13:30

GetLength returns the following:

On ASCII strings: The number of BYTES which happens to be the number of characters.

On MBCS strings: The number of BYTES which is NOT the same as the number of characters. A character that requires two BYTES will contribute two to the total of GetLength.

On UNICODE strings: The number of WCHARs which is NOT the same as the number characters (UCS-4).

Tim Smith

I'm going to patent thought. I have yet to see any prior art.

Re: CString::GetLength(), what does it in fact return?

misterbear15-Aug-05 20:05

misterbear

15-Aug-05 20:05

So if I want to be sure and in the end come out with a measure of the number of bytes (since I use this in conjunction with a filename buffer for a custom multiselect file dialog) what can I do? Is there some other way of finding out the length in bytes of a string?
In a compile-mode independant manner? I mean it seems clear that the object itself contains a buffer so it should also internally have a count of the bytes. Is this hidden? Because this behaviour from GetLength() seems to be a bit confusing...
(or maybe I am what is confused?)

firewall by-passing

avi-wildthing15-Aug-05 7:02

avi-wildthing

15-Aug-05 7:02

hello 2 all,
for my project i need to build a firewall by-passer for windows (2k & above) based on exploiting port 80 (HTTP encapsulation). the session will be opened with an external decapsulating server that will forward the original message, get the response & send it back to the source.
first:
does anyone knows an application that does that (with open source)?
second:
i think my first thing to do is to capture the outgoing packet & change it's destination ip & port while saving the original data inside the packet.
so, how can i do that, change fields etc..
third:
i wellcome any other suggetions on ways to implement this project.

p.s:
i read a lot of the things this site has to offer, but it's all very confusing & since i'm a begginer i would appriciate any help.

thanks.

Re: firewall by-passing

Rage15-Aug-05 7:04

Rage

15-Aug-05 7:04

Why would one want to bypass a firewall, especially on port 80 ?

~RaGE();

Re: firewall by-passing

avi-wildthing15-Aug-05 7:40

avi-wildthing

15-Aug-05 7:40

because the firewall allowes only http, but what if you want ftp or other applications? how do you by-pass the firewall restriction?

Re: firewall by-passing

Bob Stanneveld15-Aug-05 10:23

Bob Stanneveld

15-Aug-05 10:23

You don't!

Bypassing a firewall is a bad idea in most cases. One way to do such a thing is to write a low level driver that sends the data itself. This means that you have to write a driver on a lower level than the driver that is hooked by the firewall.

I think that the best way to accomplish the bypass of the firewall restriction, is to use an unusual port number and to state explicitly that any firewall or router should forward that port.

Bypassing the firewall doesn't work only temporary (the vendors of the software will fix the problem one day), but I'm sure that clients of your program won't like that feature. Especially changing the IP packet and sending it to another location...

Behind every great black man...
... is the police. - Conspiracy brother

Blog[^]

Re: firewall by-passing

avi-wildthing15-Aug-05 15:53

avi-wildthing

15-Aug-05 15:53

first of all, thanks!
second, i'm not trying to bypass the firewall that is on the same host as the application. i'm trying to bypass the firewall that is on the LAN edge, probably on a router (my university firewall to be exact). i can't tell if it's a regular firewall or is it combind with a proxy but i have to make it work anyway.
third, i know i have to write a driver & i searched for samples/tutorials in many websites, including pcausa.com & msdn. but because i'm new to this, i probably missed it, cause i know it's there somewhere.

Re: firewall by-passing

ThatsAlok15-Aug-05 18:27

ThatsAlok

15-Aug-05 18:27

avi-wildthing wrote:
nd, i'm not trying to bypass the firewall that is on the same host as the application

If you fireWall Allow you to communicate on Port 80 without problem, Then why not you simply use available WinInet Api! / Wrappers

"Opinions are neither right nor wrong. I cannot change your opinion. I can, however, change what influences your opinion." - David Crow

cheers,
Alok Gupta
VC Forum Q&A :- I/ IV

Re: firewall by-passing

fat_boy19-Aug-05 4:42

fat_boy

19-Aug-05 4:42

ndis protocol driver is what you need.

Look at the windows 2003 DDk. There are samples there.

For out going packets you will need to change the TCP/UDP port address to the one you want, ie, the one blocked by the firewall.

For incomming packets change it back to the allowed port.

Make sure you driver sits between the NCI driver and tcpip.sys.

Now, to write the driver. Its a pain in the arse. It will take you about 6 months.

You will need to have 2 pcs to do the debugging over a serial link, to fix the 1000 blue screen your code will generate!

But, one you get it working it is very satisfying!

Nunc est bibendum!

Hooking

celllllllll15-Aug-05 6:27

celllllllll

15-Aug-05 6:27

Hi,

Does anyone know how to use hooking ?
For example: I think we can set up a hook to detect when a window of MSn messenger closes or starts.

Thanks

Preeti9

Re: Hooking

Rage15-Aug-05 6:51

Rage

15-Aug-05 6:51

My favorite article by the wonderful Joseph M. Newcomer :

http://www.codeproject.com/dll/hooks.asp[^]

~RaGE();

Re: Hooking

celllllllll15-Aug-05 11:16

celllllllll

15-Aug-05 11:16

Can you please suggest me if hooking is a good approach to detect when a window starts and when it closes ?

Thanks

Preeti9

Re: Hooking

Rage15-Aug-05 20:34

Rage

15-Aug-05 20:34

Preeti9 wrote:
Can you please suggest me if hooking is a good approach to detect when a window starts and when it closes ?

Yes I do think so. It is actually the only way to detect in the background if a window starts or closes.

~RaGE();

Re: Hooking

celllllllll16-Aug-05 7:15

celllllllll

16-Aug-05 7:15

Thanks for the reply...

What if I use IsWindow function and then use while loop and at the end give sleep of like 1/2 a second....
Here is my program:

int APIENTRY WinMain(HINSTANCE hInstance,
HINSTANCE hPrevInstance,
LPSTR lpCmdLine,
int nCmdShow)

{
CTouchScreenAPI* pTouch;
// HWND hCalWin;
// DWORD SYNCRONIZE;
// DWORD procId;
HANDLE hProcess;
DWORD dwWaitResult;

pTouch = CTouchScreenAPI::Create();
if (pTouch)
{
if (pTouch->Init())
pTouch->Recalibrate();
else
MessageBox(NULL, "Created touchscreen interface OK, but Init() failed", "Info", MB_OK);

HWND hCalWin = ::FindWindow(pszABSPOINT_WND_CLASS_NAME, NULL);

//GetWindowThreadProcessId(hCalWin, &procId);
// retrieves the identifier of the thread that created this window.
// hProcess = OpenProcess(SYNCHRONIZE, FALSE, procId); //returns a handle to an existing process object

// Create top window class
//WinClass wc(ID_MAIN, hInstance,MainWndProc);
//hCalWin = wc.GetRunningWindow ();

// Is there a running instance of this program?
if (hCalWin !=0)
{
if( IsWindow(hCalWin) )
GetWindow(hCalWin, nCmdShow ); // handle of original window // relationship flag
//hCalWin->ShowWindow(SW_SHOW);

ShowWindow (hCalWin, SW_RESTORE);
while (1)
{
Sleep (1000);
}

return 0;
}

while (1)
{
dwWaitResult = MsgWaitForMultipleObjects(1, &hProcess, FALSE, 30*1000, QS_ALLINPUT);
if (dwWaitResult == WAIT_OBJECT_0 + 1)
{
MSG xMsg;

while (PeekMessage(&xMsg, NULL, 0, 0, PM_REMOVE))
{

if (xMsg.message == WM_QUIT)
return 1;

TranslateMessage(&xMsg);
DispatchMessage(&xMsg);
}// end of PeekMessage while loop.

int i=1;
while (i<10000)
{
i=i++;
}
}
else if (dwWaitResult == WAIT_TIMEOUT)
{
SendMessage(hCalWin, WM_CLOSE, 0, 0);
//TerminateProcess(hProcess, 0);

break;
}
else dwWaitResult == WAIT_OBJECT_0;
{
//Calibration process completed;
break;
}
}

}
else
{
MessageBox(NULL, "Unable to get pointer to Touchscreen interface", "Info", MB_OK);
}

return 0;

}

Help
Thanks

Preeti9

Re: Hooking

celllllllll15-Aug-05 11:32

celllllllll

15-Aug-05 11:32

Also, I am getting an error in this....unresolved external....
Do I need to make this application a console one ?

Thanks

Preeti9

Uninitialized pointer

valerie9915-Aug-05 5:58

valerie99

15-Aug-05 5:58

morning, all

I am looking at this code
CB_DBF_NTX* const p = OpenDbfNtxHelper( szDbfName, szIndexName, szIndexKey );
the P has value 0XCCCCCCC, if I do F10, it goes to
CB_DBF_NTX* pDbf = new CB_DBF_NTX( const_cast<char*>( szDbfName ) );
pDbf also has value 0xCCCCCCC.
does it mean p and pDbf BOTH need to be initialized or just pDbf?
thanks!

Re: Uninitialized pointer

David Crow15-Aug-05 6:27

David Crow

15-Aug-05 6:27

0xCCCCCCCC indicates an uninitialized local variable. If OpenDbfNtxHelper() fails, does it return a NULL pointer?

"One must learn from the bite of the fire to leave it alone." - Native American Proverb

Re: Uninitialized pointer

valerie9915-Aug-05 6:40

valerie99

15-Aug-05 6:40

thanks for your reply.

it will be returned void. and the OpenDbfNtxHelper will returned the valid adress.

so "0xCCCCCCCC indicates an uninitialized local variable", means oxCCCCCCC doesn't necessary cause moemory leak? thanks again!

Re: Uninitialized pointer

Maximilien15-Aug-05 7:12

Maximilien

15-Aug-05 7:12

valerie99 wrote:
so "0xCCCCCCCC indicates an uninitialized local variable", means oxCCCCCCC doesn't necessary cause moemory leak? thanks again!

never make that assumption.

every time you declare a pointer assign NULL to it; then you will be in better position to debug those kind of problems ( for example by using ASSERT )

Maximilien Lincourt
Your Head A Splode - Strong Bad

Re: Uninitialized pointer

Rage15-Aug-05 7:03

Rage

15-Aug-05 7:03

In this case

valerie99 wrote:
CB_DBF_NTX* const p = OpenDbfNtxHelper( szDbfName, szIndexName, szIndexKey );

the function probably fails and returns no pointer, but NULL.

In this case

valerie99 wrote:
CB_DBF_NTX* pDbf = new CB_DBF_NTX( const_cast( szDbfName ) );

pDbf should be a pointer on a new CB_DBF_NTX object, but the init also fails and returns NULL.

So the fact that both p and pDbf have a value of 0xCCCCCCC only shows that both pointers have the value NULL and are unititialized, but not that they are linked together in any way.

So the anwser to
valerie99 wrote:
does it mean p and pDbf BOTH need to be initialized or just pDbf?

is "Yes, both p and pDbf need to be initialized".

Check why your function OpenDbfNtxHelper and why the constructor CB_DBF_NTX fail (this has probably to do with szDbfName, which is common to both of the statements).

~RaGE();

Re: Uninitialized pointer

Maximilien15-Aug-05 7:06

Maximilien

15-Aug-05 7:06

remove the const, and do it in multiple instructions ( just for test ).

CB_DBF_NTX* p = NULL;
p = OpenDbfNtxHelper( szDbfName, szIndexName, szIndexKey );
ASSERT( p );

If p after the call to OpenDbfNtxHelper is still 0xffffffff check that the input to the function are valid.

Also, ALWAYS check that the pointers are valid, (before and) after an allocation with ASSERT.

usually, new will not fail.

are you certain you are running in DEBUG mode ?

Maximilien Lincourt
Your Head A Splode - Strong Bad

Re: Uninitialized pointer

Jose Lamas Rios15-Aug-05 9:49

Jose Lamas Rios

15-Aug-05 9:49

valerie99 wrote:
I am looking at this code
CB_DBF_NTX* const p = OpenDbfNtxHelper( szDbfName, szIndexName, szIndexKey );
the P has value 0XCCCCCCC, if I do F10, it goes to

If you look at the value of p, while the instruction pointer is at this line, then you are seeing the value before this line has executed; the value of p at this time is irrelevant. You need to look at the value of p after pressing F10

--
jlr
http://jlamas.blogspot.com/[^]

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