Skip site navigation (1) Skip section navigation (2)

Re: More strict bind param count checking

From: Ludek Finstrle <luf(at)pzkagis(dot)cz>
To: Hiroshi Inoue <inoue(at)tpf(dot)co(dot)jp>
Cc: pgsql-odbc(at)postgresql(dot)org
Subject: Re: More strict bind param count checking
Date: 2005-12-15 05:12:56
Message-ID: 20051215051256.GA21336@soptik.pzkagis.cz (view raw or flat)
Thread:
Lists: pgsql-odbc
> Hi Ludek, thanks to your recent effort to improve the psqlodbc driver.
> Please let me comment on your patch a little.

Hello,

  your comments are welcome.

> > ! 		if (ipdopts->allocated == marker_count)
> 
> Why are you using the operator '==' not '>=' ?
> AFAIK there's no such limitation.

We have problems with Visual FoxPro. It calls SQLCancel without
FreeStmt(SQL_RESET_PARAMS). It leads to failure becouse next
ExecDirect with fewer parameters in some cases think that it
has data_at_exec (which was parameters from previous).

So I think this could be good prevent againist driver failure.

> > + 	if (ipdopts->allocated != marker_count)
> 
> Why are you using '!=' not '<' ?

It's the same as previous (only negated).

> > + 	ipdopts = SC_get_IPDF(stmt);
> > + 	if (ipdopts->allocated != marker_count)
> 
> The same comment as the previous one.
> In addtion, the check is not appropriate because Exec_with_parameter_resolved
>  could be called with insufficient parameters.

I take a look at it again.

Thanks for comments

Luf

In response to

Responses

pgsql-odbc by date

Next:From: Anoop KumarDate: 2005-12-15 05:28:42
Subject: Re: Next development steps?
Previous:From: Hiroshi InoueDate: 2005-12-15 04:16:45
Subject: Re: More strict bind param count checking

Privacy Policy | About PostgreSQL
Copyright © 1996-2014 The PostgreSQL Global Development Group