Re: Transcoding problem with wine and unixodbc

Thank you. I comment.

On 04/30/2014 04:10 PM, Adrian Klaver wrote:
> On 04/30/2014 12:23 AM, Pierre Couderc wrote:
>>
>> On 04/26/2014 01:40 AM, Adrian Klaver wrote:
>>> On 04/25/2014 04:48 AM, Pierre Couderc wrote:
>>>> I am still trying to excute a MFC application (working with psqlodbc
>>>> under WinXP) under wine, unix psqlodbc, and unixodbc.
>>>>
>>>> My application is in ANSI, but the DB is in Unicode, so under
>>>> Windows, I
>>>> use Postgrsql Unicode driver (which transcodes..?) without any
>>>> problem.
>>>> Under wine, I have tried the 2 drivers (psqlodbcw.so and psqlodbca.so)
>>>> and with the same incorrect result.
>>>
>>> Hmm. I would have expected different results.
>> Me too.
>>>
>>> Are you sure your program actually picked up the change in drivers?
>> Yes, I am sure : I choose "PostgreSQL ANSI" (in odbc.ini) I get a good
>> result but the bad <Ã"> result. I change odbc.ini to a unexistant
>> driver : I get an application error, then I choose "PostgreSQL Unicode",
>> and I get the same bad <Ã"> result.
>>>
>>>> For exemple <è> becomes <Ã"> (in 2 characters).
>>>
>>> Can you get a psql connection to the database?
>> yes.
>>>
>>> And if so what does it show?
>>>
>> I do not know (as I do not know how to use pgsql). But I am sure that
>> the database is Unicode, and for exmple PGadmin III (either on the same
>> debian machine or on W7) reads it without problem.
>
> Once you have psql open, type the query you want:
>
> select * from some_table;
>
> where some_table is the one with the characters you are having
> problems with.
>
As expected by me, as pgAdminIII, psql prints correctly.
>
>>> Also what exactly is the encoding your client is using?
>>>
>> Mmm, I am not sure to know what is "my" Client. I suppose that the
>> client of the DB is psqlodbc.
>
> Well psqlodbc is more of an adapter that is passing information from
> your application to Postgres.
>
>> Or if you want to speak of my application, it is under wine in ANSI.
>
> ANSI is a general term. What is important is what character
> set(encoding) you are actually using in your application.
>
LATIN1 (ISO-8859-1),
>> And I do not know who is responsible for translation, normally it should
>> be the unixodbc or psqlodbc..