[unixODBC-dev] SQLEndTran returns SQL_SUCCESS forHY010(Function sequence error)

Nick Gorham nick at lurcher.org
Wed Jun 17 12:22:18 BST 2009


G, Hemanth Kumar wrote:
> Yes, it is resolved. Thanks Nick :-)
>
> I have one more doubt. 
> Some of the ODBC calls are trying to change the odbc.ini to insert some
> index or root information.
> For e.g - SQLExecDirect(hstmt, (SQLCHAR *)"create unique index unq on
> tmp(f1)", SQL_NTS);
> Ultimately, the odbc.ini file has extra tabs and spaces added.
> Here onwards, all the subsequent calls fail whenever they try to read
> the information from odbc.ini.
>
>   
Hmm, not sure why that call would affect the ini file.
> I found out the line of code from DriverManager which is doing this
> modification.
> _iniDump.c#74
> uo_fprintf( hStream, "%s\t\t%c %s\n", hIni->hCurProperty->szName,
> hIni->cEqual, hIni->hCurProperty->szValue );
>
> I am not very much sure about the standard format of odbc.ini. 
> Should it be of this format which you are using with two tabs and a
> space.
>   
There isn't one AFAIK, windows doesn't use a ini file now anyway. I did 
toy with using ldap as a option to replace the ini, and also moving the 
ini to XML. But my take is that from the apps and drivers perspective it 
should be hidden behind the API calls that alters the information store, 
SQLGetPrivateProfileString, SQLPutPrivateProfileString etc.

We tried to make the ini code in unixODBC read other DM's formats, but I 
can't say what they do.

-- 
Nick


More information about the unixODBC-dev mailing list