Andrea Aime wrote:
> I begin to see some light now that I've spent some hours looking
> at the logs and at the source code. Autocad first prepares a
> statement with SQLPrepare, and the statement looks like:
> "SELECT CLASSE FROM tav974 WHERE SIAPK974=?"
> Then it calls SQLNumResultCols, which in turn calls SC_pre_execute
> and finally SQLExecute wich fails because there is still no
> parameter binding. If I set "Parse statements = 1" then SQLNumResultCols
> runs fine and returns 1, but the subsequent SQLDescribeCol fails
> because it calls SC_pre_execute... That happens in the following code
> (from results.c):
> if (stmt->parse_status != STMT_PARSE_FATAL && stmt->fi &&
> if (icol >= stmt->nfld)
> stmt->errornumber = STMT_INVALID_COLUMN_NUMBER_ERROR;
> stmt->errormsg = "Invalid column number in DescribeCol.";
> SC_log_error(func, "", stmt);
> return SQL_ERROR;
> mylog("DescribeCol: getting info for icol=%d\n", icol);
> fieldtype = stmt->fi[icol]->type;
> col_name = stmt->fi[icol]->name;
> precision = stmt->fi[icol]->precision;
> mylog("PARSE: fieldtype=%d, col_name='%s', precision=%d\n", fieldtype,
> col_name, precision);
> if (fieldtype > 0)
> parse_ok = TRUE;
> * If couldn't parse it OR the field being described was not parsed
> * (i.e., because it was a function or expression, etc, then do it the
> * old fashioned way.
> if (!parse_ok)
> SC_pre_execute(stmt); <======= HERE!!!!!!!
> Now, it seems that this happens because stmt->fi[icol] doesn't get
> properly initialized...
Isn't the second parameter of the SQLDescribeCol() zero ?
In response to
pgsql-odbc by date
|Next:||From: Hiroshi Inoue||Date: 2001-04-20 01:04:40|
|Subject: Re: How to configure iodbc access to local postgres db?|
|Previous:||From: Rob Yampolsky||Date: 2001-04-19 17:50:56|
|Subject: How to configure iodbc access to local postgres db?|