Re: Fwd: Re: unicode in 7.1 - Mailing list pgsql-general

No this problem was fixed when I rebuild the database with UNICODE encoding.  No more '?'.  Now the data just gets *truncated* when it hits a non-Latin character. 

Here is my classpath:

/usr/jdk1.3/jre/lib/rt.jar:/usr/jdk1.3/lib/tools.jar:/usr/jakarta-oro-2.0.1/jakarta-oro-2.0.1.jar:/usr/jakarta-regexp-1.2/jakarta-regexp-1.2.jar:/usr/jakarta-servletapi-3.2/lib/servlet.jar:/usr/jakarta-tomcat-3.2.1/lib/jasper.jar:/usr/jakarta-tomcat-3.2.1/lib/webserver.jar:/usr/Jetspeed-1.1/Jetspeed.jar:/usr/cocoon-1.8/lib/stylebook-1.0-b2.jar:/usr/cocoon-1.8/bin/cocoon.jar:/usr/xerces-1_2_0/xerces.jar:/usr/jakarta-ant/lib/ant.jar:/usr/jakarta-ant/lib/jaxp.jar:/usr/jakarta-ant/lib/optional.jar:/usr/jakarta-ant/lib/parser.jar:/usr/xalan_1_2_D01/xalan.jar:/usr/xalan_1_2_D01/bsf.jar:/usr/xalan_1_2_D01/bsfengines.jar:/usr/aoserv/lib/acme.jar:/usr/aoserv/lib/aocode-public.jar:/usr/aoserv/lib/aoserv-private.jar:/usr/aoserv/lib/damud.jar:/usr/aoserv/lib/danshome.jar:/usr/castor/castor-0.8.8.jar:/usr/cos/lib/cos.jar:/usr/ecs-1.3.3/ecs-1.3.3.jar:/usr/freemarker152/freemarker-jdk1.2.jar:/usr/gnu.regexp-1.0.8/lib/gnu-regexp-1.0.8.jar:/usr/jaf-1.0.1/activation.jar:/usr/jakarta-slide-1.0m4/slide/lib/webdav.jar:/usr/jakarta-slide-1.0m4/slide/lib/slide.jar:/usr/javachart/jars/kcServlet.jar:/usr/javamail-1.1.3/mail.jar:/usr/jdbc-2.0/jdbc-2.0.jar:/usr/jsse1.0.2/lib/jcert.jar:/usr/jsse1.0.2/lib/jnet.jar:/usr/jsse1.0.2/lib/jsse.jar:/usr/Apache-Jyve-20000907/src/java:/usr/mm.mysql-2.0.4/mm-mysql.jar:/usr/openxml/openxml-1.2.jar:/usr/ozone-0.5.5:/usr/share/pgsql/jdbc7.1-1.2.jar:/usr/pop3-1.1.1/pop3.jar:/usr/soap-2_0/lib/soap.jar:/usr/spfc-0.2.0/bin/spfc.jar:/usr/Apache-Turbine-20000907/src/java:/usr/village-1.3/village-1.3.jar:/usr/webmacro/webmacro.jar:/usr/xang_0_0_2/fesi.zip:/usr/xang_0_0_2/xang.jar:/usr/xmlrpc-java/lib/xmlrpc.jar:/usr/ApacheJSSI-1.1.2/src/java/ApacheJSSI.jar:/usr/interclient/interclient.jar:/usr/PoolMan-1.4.1/lib/PoolMan.jar:/usr/fop-0_13_0/fop_bin_0_13_0.jar:/usr/fop-0_13_0/lib/xml.jar:/wwwgroup/culley/classes:/www/flashcard/classes:/www/banyan/classes

In the same directory as the 7.1-1.2 driver is a copy of the old driver.  Is there someway I could still be hitting the old one?

An interesting anomoly-- When I hit my remote production postgres server from my local tomcat (normally I'm using tomcat on the same machine as the database) non-English characters *were* displayed correctly.  I was through the jdbc7.0-1.2.jar driver though!  Very confusing...

I'm currently updating my local dinky little linux box with postgres 7.1.3 so my dev environment more closely matches the production.  Debugging should be easier...

culley 



At 04:33 PM 9/11/01 -0700, you wrote:

Culley,

That all looks fine.  I don't understand why you would be having problems with these settings.  Are you sure you are using the 7.1 jdbc drivers?  One other thing I can think of to check is to do a 'show client_encoding'.  This should also return UNICODE.

So to restate your problem, you get the 8bit characters returned as ?'s through jdbc, but it works correctly through psql.  Is this a correct restatement of your problem?

thanks,
--Barry

Culley Harrelson wrote:

At 11:45 AM 9/11/01 -0700, you wrote:

select getdatabaseencoding();

UNICODE

select datname, encoding from pg_database;

 flashcard     |        5
All other databases on the server had 0 for encoding.
Unfortunately my dev environment still has postgres 7.0.  I updated my connection pool object and it works on dev but not on prod with 7.1.  I really don't want to update my dev environment untill I can do the whole machine from scratch but I fear I might have to in order to debug this...
Culley

_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com