Обсуждение: handle a ECPG_bytea typo

Hi, hackers

The source looks like:

    case ECPGt_bytea:
    {
        struct ECPGgeneric_varchar *variable =
        (struct ECPGgeneric_varchar *) (var->value);

        ......
    }

I think the developer intend to use struct ECPGgeneric_bytea instead of struct ECPGgeneric_varchar

Is this thoughts right?

I have wrote a patch to fix this typo

On Fri, Jul 24, 2020 at 1:35 PM Wang, Shenhao
<wangsh.fnst@cn.fujitsu.com> wrote:
>
> Hi, hackers
>
> The source looks like:
>
>         case ECPGt_bytea:
>         {
>                 struct ECPGgeneric_varchar *variable =
>                 (struct ECPGgeneric_varchar *) (var->value);
>
>                 ......
>         }
>
> I think the developer intend to use struct ECPGgeneric_bytea instead of struct ECPGgeneric_varchar
>
> Is this thoughts right?
>
> I have wrote a patch to fix this typo

I felt the changes look correct. The reason it might be working
earlier is because the structure members are the same for both the
data structures ECPGgeneric_bytea & ECPGgeneric_varchar.

Regards,
Vignesh
EnterpriseDB: http://www.enterprisedb.com

On Sat, Jul 25, 2020 at 07:22:15AM +0530, vignesh C wrote:
> I felt the changes look correct. The reason it might be working
> earlier is because the structure members are the same for both the
> data structures ECPGgeneric_bytea & ECPGgeneric_varchar.

ECPGset_noind_null() and ECPGis_noind_null() in misc.c show that
ECPGgeneric_bytea is attached to ECPGt_bytea.  The two structures may
be the same now, but if a bug fix or a code change involves a change
in the structure definition we could run into problems.  So let's fix
and back-patch this change.  I am not spotting other areas impacted,
and I'll try to take care at the beginning of next week.
--
Michael

On Sat, Jul 25, 2020 at 06:17:42PM +0900, Michael Paquier wrote:
> ECPGset_noind_null() and ECPGis_noind_null() in misc.c show that
> ECPGgeneric_bytea is attached to ECPGt_bytea.  The two structures may
> be the same now, but if a bug fix or a code change involves a change
> in the structure definition we could run into problems.  So let's fix
> and back-patch this change.  I am not spotting other areas impacted,
> and I'll try to take care at the beginning of next week.

Okay, fixed as e971357.  The issue came from 050710b, so this fix was
only needed in 12~.
--
Michael