why gSOAP only require the root CA when accessing HTTPS? - C / C++ / MFC Discussion Boards

Richard Andrew x6425-Sep-13 18:21

25-Sep-13 18:21

Instead of updating your application with an MSI, you're supposed to create a "patch". The MSI is only for installing, repairing and removing the application.

If you are only distributing an update, then use a patch instead.

What software are you using to author your MSI's?

The difficult we do right away...
...the impossible takes slightly longer.

Re: software update solution

koll Zhu25-Sep-13 20:32

koll Zhu

25-Sep-13 20:32

there'r 2 cases. one maybe means a "patch" for adding some more resources.
in the other case , i need an entirely new version. my question is about the second case.
so i want to take a solution of reinstallation,though, im not sure whether its resonable or not...
and i cant understand whats ur mean about authoring my msi's.

Re: software update solution

Richard Andrew x6426-Sep-13 2:39

Richard Andrew x64

26-Sep-13 2:39

Try running the installer with the "\quiet" option. This prevents any user interface.

See: Installer Command Line Options[^]

The difficult we do right away...
...the impossible takes slightly longer.

Re: software update solution

koll Zhu26-Sep-13 14:48

koll Zhu

26-Sep-13 14:48

yes,it works.there's also a doubt why i cant run the uninstall .exe if it has a different productcode from setup.msi.
and how can i get the productcode.
thank u for ur reply,its very useful

Re: software update solution

Richard Andrew x6426-Sep-13 17:34

Richard Andrew x64

26-Sep-13 17:34

If you need to get the product code from the MSI, you could use the tool called Orca.

It's part of the Windows SDK: Orca[^]

You can see the contents of the MSI file with this tool.

The difficult we do right away...
...the impossible takes slightly longer.

Re: software update solution

Richard Andrew x6426-Sep-13 11:09

Richard Andrew x64

26-Sep-13 11:09

You might want to look into this function:

MsiSetInternalUI[^]

The difficult we do right away...
...the impossible takes slightly longer.

Re: software update solution

koll Zhu26-Sep-13 15:06

koll Zhu

26-Sep-13 15:06

yeah

why gSOAP only require the root CA when accessing HTTPS?

szh12125-Sep-13 17:10

szh121

25-Sep-13 17:10

I have a self-signed CA Root. An intermediate certificate was signed by this CA Root, and my server is the 3rd level which is signed by the intermediate certificate. What i have is like:

CA Root --> Intermediate Cert --> Server Cert

What i'm using is the function soap_ssl_client_context() to connect to the server and calling its web services. The function soap_ssl_client_context() requires one certificate file with PEM format as the cert file of the server.

The fact i'm facing is that, if i only put the CA Root certificate only (without the intermediate cert and server cert) in the PEM file, all the connection and web service calling are succeed. Is that reasonable? I mean it can pass the verification if only CA Root was provided. Is there anyway to make it verify the full certificate chain instead just the CA root?

The code looks like bellow:

C++

try 
{
    soap_ssl_init();
    soap_init(soap);

    if (soap_ssl_client_context(
        soap, 
        SOAP_SSL_DEFAULT,
        NULL,   /* keyfile: required only when client must authenticate to server (see SSL docs on how to obtain this file) */ 
        NULL,    /* password to read the key file (not used with GNUTLS) */ 
        "C:\\certfiles.pem", /* cacert file to store trusted certificates (needed to verify server) */              
        NULL,  /* capath to directory with trusted certificates */ 
        NULL     /* if randfile!=NULL: use a file with random data to seed randomness */ 
        ))
    { 
        isSucceed = false;
    }

} catch (...) {
    isSucceed = false;
}

Any comment will be appreciated. Thanks.

Re: why gSOAP only require the root CA when accessing HTTPS?

Richard MacCutchan25-Sep-13 21:39

Richard MacCutchan

25-Sep-13 21:39

I don't think this is a C++ programming issue, you would be better posting at http://www.codeproject.com/Forums/1640/Web-Development.aspx[^].

Veni, vidi, abiit domum

how to make this working: srand(time(NULL));

mrby12325-Sep-13 12:58

mrby123

25-Sep-13 12:58

I do not know why?
one program works with srand(time(NULL));
, but the other one not working.

What do I need to make srand(time(NULL)) work ?

Please help.

Re: how to make this working: srand(time(NULL));

Richard Andrew x6425-Sep-13 13:12

Richard Andrew x64

25-Sep-13 13:12

You have to define what you mean by "not working."

Is there a crash? Do you have an error code? What goes wrong?

The difficult we do right away...
...the impossible takes slightly longer.

CRegKey read from registry?

bosfan25-Sep-13 6:19

bosfan

25-Sep-13 6:19

Hi,
i need some help with CRegKey to read from registry Frown | :(

Here what works fine:

<br />
int *target = NULL;<br />
AfxGetApp()->GetProfileBinary(m_strKey, _T("key"), (BYTE **)&target, &bytes);<br />

but i need to do this with registry key, like:

<br />
// CRegKey is there and opened without errors<br />
int *target = NULL;<br />
myregkey.QueryBinaryValue(_T("key"), (BYTE**)&target, &bytes);<br />

Error is allways ERROR_MORE_DATA what means they are more information to read as i do?
To write binary data to registry works fine to, only read is my problem now.
Only for example here how i write binary to registry:

<br />
int *value = new int[count];<br />
myregkey.SetBinaryValue(_T("key"), (BYTE*)value, sizeof(int) * count);<br />

Can somebody help me how to read binary from registry with CRegKey::QueryBinaryValue??

Thanks in advance
bosfan

Re: CRegKey read from registry?

Richard Andrew x6425-Sep-13 7:47

Richard Andrew x64

25-Sep-13 7:47

ERROR_MORE_DATA usually means your buffer is not large enough. Call the function with a buffer that is large enough to receive the binary value.

The difficult we do right away...
...the impossible takes slightly longer.

Re: CRegKey read from registry?

bosfan25-Sep-13 21:46

bosfan

25-Sep-13 21:46

Hi,
thanks for answer, yes the buffer was wrong and i corrected this
and now works.

regards
bosfan

Re: CRegKey read from registry?

Richard MacCutchan25-Sep-13 21:37

Richard MacCutchan

25-Sep-13 21:37

As per the notes for the RegQueryValueEx function[^]

Quote:
If the buffer specified by lpData parameter is not large enough to hold the data, the function returns ERROR_MORE_DATA and stores the required buffer size in the variable pointed to by lpcbData. In this case, the contents of the lpData buffer are undefined.

So set your variable bytes to zero and make the call, then when it returns it will contain the required size. You then allocate a buffer of the required size and repeat the call.

Veni, vidi, abiit domum

Re: CRegKey read from registry?

bosfan25-Sep-13 21:45

bosfan

25-Sep-13 21:45

Hi,
thanks for Answer, you right, i solve it in this way!
I didn know that i can check "the required size" Frown | :(

Now:

<br />
srg.QueryBinaryValue(_T("key"), NULL, &bytes);<br />
// then if ERROR_SUCCESS <br />
int *iArr = new int[bytes];<br />
srg.QueryBinaryValue(_T("key"), (BYTE**)&iArr[0], &bytes);<br />

Now i now how big is the needed size, it is saved in "bytes".
And i have all Data from the registry.

Best regards
bosfan

Re: CRegKey read from registry?

Richard MacCutchan25-Sep-13 22:03

Richard MacCutchan

25-Sep-13 22:03

C++

int *iArr = new int[bytes];

Your buffer will be 4 (or even 8) times bigger than necessary; you should use:

C++

BYTE *iArr = new BYTE[bytes];

Veni, vidi, abiit domum

Re: CRegKey read from registry?

bosfan25-Sep-13 22:32

bosfan

25-Sep-13 22:32

You right again, i change this, but i think if i do this in this way i can call the function
without cast to byte, directly?
Something like this:

<br />
// regkey.QueryBinaryData(_T("key"), (BYTE**)&iArr[0], &bytes); // before<br />
regkey.QueryBinaryData(_T("key"), &iArr[0], &bytes); // now with byte array?<br />

i try this and this works to, is there any thing wrong now?

bosfan

Re: CRegKey read from registry?

Richard MacCutchan25-Sep-13 22:45

Richard MacCutchan

25-Sep-13 22:45

I cannot find a method called QueryBinaryData so I presume you are referring to CRegKey::QueryBinaryValue[^]. You could also use the simpler syntax for your buffer address thus:

C++

regkey.QueryBinaryValue(_T("key"), iArr, &bytes);

The compiler may also still insist that you cast the buffer address to a void*.

Veni, vidi, abiit domum

Re: CRegKey read from registry?

bosfan26-Sep-13 20:42

bosfan

26-Sep-13 20:42

Oh no, yes of course i mean QueryBinaryValue Frown | :(

my mistake Cry | :((

(.
One question now, is there any difference in Unicode Version, my App is in Unicode?

Regards
bosfan

Re: CRegKey read from registry?

Richard MacCutchan26-Sep-13 21:30

Richard MacCutchan

26-Sep-13 21:30

If the data you are extracting is binary, then no, there is no difference.

Veni, vidi, abiit domum

Socket Communication - Server with 2 different clients

manoharbalu25-Sep-13 0:47

manoharbalu

25-Sep-13 0:47

I have a client and server MFC applications that uses CSocket. The Server exchanges data with the client 1 through port 1500.

I want to create a new client application entirely different from client 1 which needs to connect to the server but needs entirely different data than that of client 1 which already connects to the Server in Port 1500.

Please suggest me some technical details in terms of port and way of establishing a unique connection with the server.

In case If I connect the 2 clients to the same port where the server is listening, how can I know from the Server code that application (client 1) is requesting a connection and client 2 is requesting a connection.

Please let me know how can I distinguish applications(clients) from Server side?

Re: Socket Communication - Server with 2 different clients

Richard MacCutchan25-Sep-13 1:01

Richard MacCutchan

25-Sep-13 1:01

Please don't use bold font in your questions, it is considered rude. Also we are all perfectly capable of reading a clearly written question.

As to your problem, there is really nothing much to do other than designing a simple protocol for your clients:
On connect the server starts a new thread which will talk to that client.
The client then sends a request which indicates which data it requires
The server gets the data and sends to the client.

Veni, vidi, abiit domum

Re: Socket Communication - Server with 2 different clients

manoharbalu25-Sep-13 2:24

manoharbalu

25-Sep-13 2:24

First of all, I apologize for the inconvenience regarding my font usage and I assure that I won't continue that henceforth.

I want to add something to my above question.

1. The Server and client_1 are already developed applications which communicate flawlessly.
2. There may be a multiple client instances of client_1 that may run and get connected to the server and exchange data. So if any of the data changes in the server, the same gets reflected in all of the client instances of client_1.
3. Now I want to add on a new client_2 entirely different from client_1. So any number of client_2 instances may run and connect to the same server to exchange some other set of data.
4. The Server is listening in Port 1500 for eg. Suppose it gets a client request, now how am I from the server side going to recognize the client before calling accept() so that I could route the client accordingly to exchange data.
For Eg: If I know that it is from client_1 I will call SendDataClient1() in a thread. or I will call SendDataClient2() if I know that it is from client_2
Please suggest me with some code to find out a way to differentiate the client_1 and client_2 for the server

Re: Socket Communication - Server with 2 different clients

Richard MacCutchan25-Sep-13 2:59

Richard MacCutchan

25-Sep-13 2:59

The only way to distinguish, would be to have a listener on a different port. So, client1 types connect to 1500, and client2 types connect to 1501 (for example).

Veni, vidi, abiit domum

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