encoding problem

Hello,

I am testing Win2k clients in reading/writing arabic in Postgres DB via web
(apache with scripts written in PHP).

I have 2 PostgreSQL DB created with -E unicode:
One on RH linux 7.1 (postgresql-7.0.3-8) (CASE 1)
The other on RH linux 8.0 (postgres-7.2.2-1)  (CASE 2)

In case 1, a win2k client writes to a postgres7.0 DB and the arabic letters
can clearly be read in "psql" and on the web.

In case 2, a win2k client writes to the 7.2 postgres, the arabic input is in
psql:

test=# select * from ttest;
  nom_fr   |        nom_ar
-----------+-----------------------
nathalie  |                          //invisible when no charset defined in
web page
nathalie  |
nathalie1 |
latino    |
dfww      | بيب   //has this format when charset=UTF8
dfww      | بيب
(6 rows)


Is that related to the automatic conversion in postgres7.2?

When i get the &#xxxx; format, each letter is consuming 7 caracters in the
database!

I cannot understand the difference btw the 2 versions of postgres.
I would greatly appreciate any help.
and what should i do to see the arabic letters in the version 7.2?

Thank you in advance.

Regards
Nathalie





_________________________________________________________________
Add photos to your messages with MSN 8. Get 2 months FREE*.
http://join.msn.com/?page=features/featuredemail