C / C++ / MFC

19-Nov-02 22:12

I keep getting an unknown runtime error 3705 or 0x0800A0E79 when I tried to open the recordset.

I tried look them up with VC's "Error Lookup". It says "Message Not Found"...

// ADOSchema.cpp : Defines the entry point for the console application.
//

#include "stdafx.h"

#define INITGUID
#import "C:\Program Files\Common Files\System\ADO\msado15.dll" rename_namespace("ADOCG") rename("EOF", "EndOfFile")
using namespace ADOCG;
#include "icrsint.h"

void DisplayHR(HRESULT hr, char * lpszErrorMsg)
{
int nCode = -1;

cout << endl;

cout << "ERROR MSG: " << lpszErrorMsg << endl;

nCode = SUCCEEDED(hr);
cout << "SUCCEEDED(hr) = " << nCode << endl;

nCode = HRESULT_SEVERITY(hr);
cout << "HRESULT_SEVERITY(hr) = " << nCode << endl;

nCode = HRESULT_CODE(hr);
cout << "HRESULT_CODE(hr) = " << nCode << endl;

return;
}

int main(int argc, char* argv[])
{
HRESULT hr;
char * pszErrorMsg = NULL;

char * pszTableName = NULL;
char ** pTableArray = NULL;
int nMaxTableNameLength = -1;
int nNumUserTable = -1;

::CoInitialize(NULL);

pszErrorMsg = new char[100];

//STEP 1: Connection object
_ConnectionPtr pConn = NULL;

hr = pConn.CreateInstance( __uuidof(Connection) );
if(hr!=S_OK)
{
//Additional error handling here.
:isplayHR(hr, "ERROR: failed to instantiate connection pointer");
return 1;
}

//Data Source=norm OR 127.0.0.1 (It doesnt matter, it's the default instance)
hr = pConn->Open(L"Provider=sqloledb; Data Source=127.0.0.1; Initial Catalog=realtime", L"sa", L"sa", NULL);
if(hr!=S_OK)
{
//Additional error handling here.
:isplayHR(hr, "ERROR: failed to establish connection with target database");
return 1;
}

//STEP 2: Command object.
_CommandPtr pCmd;
pCmd.CreateInstance( __uuidof(Command) );
pCmd->ActiveConnection = pConn;
pCmd->CommandText = "SELECT * FROM sysobjects";

//STEP 3: Recordset - play with it.
_RecordsetPtr pRs;

try
{
pRs = pCmd->Execute(NULL, NULL, adCmdText);
}
catch(_com_error &e)
{
//Additional error handling here.
cout << "ERROR: Failed to execute command." << endl;
cout << "ERROR msg: " << e.ErrorMessage() << endl;

hr = e.Error();
cout << "ERROR: HRESULT_CODE( e.Error() ): " << HRESULT_CODE(hr) << endl;

return 1;
}

//Null variant:
_variant_t vNull;
vNull.vt = VT_ERROR;
vNull.scode = DISP_E_PARAMNOTFOUND;

try
{
//CursorType: adOpenForwardOnly; LockType: adLockReadOnly; Option: adCmdUnknown
pRs->Open(vNull, vNull, adOpenForwardOnly, adLockReadOnly, adCmdUnknown);
}
catch(_com_error &e) //QUESTION: Runtime error HERE!!! Cant figure out why!!
{
//Additional error handling here.
cout << "ERROR: Failed to open recordset." << endl;
cout << "ERROR: e.ErrorMessage(): " << e.ErrorMessage() << endl; //CAUTION: This turns out to be "Unknown error 0x800A0E79.

hr = e.Error();
cout << "ERROR: HRESULT_CODE( e.Error() ): " << HRESULT_CODE(hr) << endl; //CAUTION: And this turns out to be 3705.

return 1;
}

//STEP 4: Retrieve a list of user tables from sysobjects
nMaxTableNameLength = 100;
pszTableName = new char [nMaxTableNameLength];

pRs->MoveFirst();
while( !EOF )
{
hr = pRs->MoveNext();
if( hr!=S_OK)
{
//Additional error handling here!
DisplayHR(hr, "ERROR: MoveNext");
}

//If field=xType==1, Then the object is a user table. Increment nNumUserTable here.

}

//STEP 4: Get individual table properties: Field/data type/number of records/number of columns/PK-FK.

//STEP 5: Cleanup
delete [] pszTableName;
delete [] pszErrorMsg;

CoUninitialize();

return 0;

}

Can anyone spot the error? i'm dying...

norm

Shaun Wilde19-Nov-02 22:23

Shaun Wilde

19-Nov-02 22:23

A quick search in google on 800A0E79 cam up with

Once you have called the Open method on an ADODB.RecordSet object, you must call the Close method before you can reuse it. Check for more than one Open using the same object name, or an Open in a loop without a matching Close

so you need to Close the pCmd object before you Open it again.

Stupidity dies.
The end of future offspring.
Evolution wins.
- A Darwin Awards Haiku

devvvy19-Nov-02 22:33

19-Nov-02 22:33

u're right, i:

pRs->Close

before I pRs->Open(...).

and it worked. all exception suddenly disappeared. The following line must have implicitly openned the recordset...

pRs = pCmd->Execute(NULL, NULL, adCmdText);

I still dont understand, i will search the error code on google and see if i can find a better explanation for this...

if "pRs = pCmd->Execute(NULL, NULL, adCmdText)" implicitly "open" the recordset, then why do we need to call pRs->Open(...)

norm

Shaun Wilde19-Nov-02 22:36

Shaun Wilde

19-Nov-02 22:36

Its the connection pConn and you had already opened it and then attached it to your pCmd

hence it was open before you called Execute.

Stupidity dies.
The end of future offspring.
Evolution wins.
- A Darwin Awards Haiku

Shaun Wilde19-Nov-02 22:38

Shaun Wilde

19-Nov-02 22:38

My ADO is pretty poor but I don't think you have to open the recordset - I think you can just get on tith reading from it after you have called Execute

Stupidity dies.
The end of future offspring.
Evolution wins.
- A Darwin Awards Haiku

devvvy19-Nov-02 22:46

19-Nov-02 22:46

thanks for the feedback. but one silly question:

what does it mean by pRs->Close()??

pConn->Close() is an API that tells SQL server to drop the connection and it is NOT equivalent to "delete pConn". That's okay and I can understand.

But recordset is a data structure. What do u mean by "closing" a data structure? remove it from memory?

I am confused by the two concepts:
1. closing... a memory/variable?
2. removing a variable from memory/stack/...etc

thanks.

norm

devvvy19-Nov-02 22:49

19-Nov-02 22:49

from MSDN:

"When you have concluded your operations over an open Recordset object, use the Close method to free any associated system resources. Closing an object does not remove it from memory; you can change its property settings and use the Open method to open it again later. To completely eliminate an object from memory, set the object variable to Nothing"

so, what about the connection object associated/attached to a recordset?

norm

RTF Formatting & preview

Steve S19-Nov-02 22:06

Steve S

19-Nov-02 22:06

I'm using a CRichEditCtrl object as one of the many windows in an app, and now need to use print preview. I have other windows, such as grids, and everything works fine except....

I'm using RT v3.0 control under NT4/2000/XP. When I ask it to format for printing, the rendering dc and formatting dc are the same (m_hDC and m_hAttribDC respectively), and it looks OK. When I format it for preview, the rendering DC and formatting dc aren't the same (which you'd expect, I guess).
However, it consistently doesn't preview the final line on the 'page' (ie in the rect I've given it. Now I know that it's a preview, but my customers expect WYSIWYG, so they'll complain about missing lines;
I'm rendering to a memory dc and blitting, but I know it's not the blit doing the clipping, as this is 'transparent' to the printing (it's done in OnPrint for every page, and the rect for printing the RTF is way smaller than the page).

Can anyone point to an RTF rendering example that works with MFC's print preview (or not, if it renders to a memory dc!) that doesn't exhibit this problem?

I'm using FormatRange and DisplayBand to do the formatting and rendering, but MSDN seems pretty light on preconditions for these, doesn't mention stuff like mapping modes needed etc, although it seems that the DC has to be MM_TEXT before it works at all.

Steve S
[This signature space available for rent]

Miguel Lopes20-Nov-02 4:51

Miguel Lopes

20-Nov-02 4:51

Im having a similar problem with previewing a RTF. Im not using FormatRange neither Display band; im using PrintInsideRect, which is working fine except for some problems not related to this.
Be carefull because there are different ways to get the total number of chars in the RichEdit. You have to be coherent between the different functions that may use the number of chars you want. For example, PrintInsideRect retrieves the amount of chars that fit in a specified are, but that amount may not be the same if you use some other function, as it may count some other chars like CR/LF, etc
That way, you final line may not appear because you're not sending enough chars.

Steve S20-Nov-02 5:04

Steve S

20-Nov-02 5:04

Thanks for that.

I don't think it applies in this case though, as I'm consistently using FormatRange to get the next character (ie starting position), and I'm getting the same values for both preview and print modes. However, it does give me a few ideas which I'll try later; if it works, I can feel an article approaching.

If only MS had documented this better, it would be far simpler.

Steve S
[This signature space available for rent]

Gary R. Wheeler20-Nov-02 14:02

Gary R. Wheeler

20-Nov-02 14:02

I've had problems with print previews in the past. Mostly what I've seen are roundoff errors in line spacing between the printed version and the screen version. Tweaking the font values for the screen version did the trick.

Software Zen: delete this;