Re: Character set blues...

From: Jeremy <newspostings_at_hazelweb.co.uk>
Date: Thu, 24 Mar 2005 14:43:06 -0000
Message-ID: <MPG.1cace24cdc08fb81989df5@news.individual.net>

In article <d1uek3$cav$1_at_news1.zwoll1.ov.home.nl>, Frank van Bortel says...
> Jeremy wrote:
> > In article <d1sdvr$7vr$1_at_news2.zwoll1.ov.home.nl>, Frank van Bortel
> > says...
> >
> >>*How* do you look inside your clob? Just an hint: could it
> >>be your *tool* to inspect your clob does not recognize the "๋" in
> >>ori๋nterend and displays a question mark?
> >>
> >
> >
> > No the tool will display fine the accented chars (have tested this
> > already). Simply put, we receive the XML into the clob as described and
> > simply insert that into a table (also defined as a clob column). No
> > messing about with the the data whatsoever - and the accented chars have
> > gone.
> >
>
> Well, in that case, I don't understand the:
> <quote>
> Whe we look into the CLOB we don't see these - instead we see e.g.
> "ori?nterend".
> </quote>

What I meant was that if I used some other input mechanism to insert the data with accented characters *instead* of doing the HTTP post then I could see the data

But for the data received via the HTTP post the data is stripped before the pl/sql procedure can see it.

>
> OK - may we conclude from this that:
> - the HTTP is correct

check
> - the interface to the database is correct
check
> - the database itself is correct

check
> - the tool(s) used to inspect the CLOB is/are correct
check
> with correct being correct as far as NLS settings are concerned.
>
> >
> >>The http post is in what character set?
> >
> >
> > Tried this via 2 DADs - one ISO-8859-1 the other UTF-8 - same result in
> > both cases.
> >
>
> See above - seems OK

In testing by me (rather than the 3rd party) I have actually discovered that posting to the ISO-8859-1 defined DAD *is* working correctly - the content of the CLOB is as intended.

<snip>

> If you still have a problem, explain where.

The basic problem is that a UTF-8 encoded input stream sent via HTTP post via a DAD configured thus:

<Location /utf8>

        SetHandler pls_handler
        Order deny,allow
        Allow from all
        AllowOverride None
        PlsqlDatabaseUsername test1
        PlsqlDatabasePassword xxxxxxxxxxx
        PlsqlDatabaseConnectString netdev9i
        PlsqlAuthenticationMode Basic
        PlsqlNLSLanguage AMERICAN_AMERICA.UTF8

</Location>

results in us not being able to see the data as intended.

We have tried this with a VARCHAR2 instead of CLOB but it made no difference.

-- 

jeremy

Received on Thu Mar 24 2005 - 08:43:06 CST