Hibernate version: 3.0

Basically after I made switch from my old ways (using prepared statements and SQL statements) to Hibernate, my client pointed out a problem:

Whenever I would do HibernateUtil.update( myclass ), where I only changed some INT value, the TEXT field containing description of an object would (apparently) also be updated (even though it wasn't changed) and some gibberish started to appear in text.

When I say gibberish I mean random out-of-normal-range ascii characters (like As with dots on top and stuff), and they seemed to appear in random spots (but normally as a first letter in the sentence).

I have no idea what could be causing this (I doubt it's encodings, since most of the stuff stays fine), and I doubt my client is putting crap in there just to annoy me.

Anybody dealt with anything like this before?

really sounds like encoding issues .... hibernate detects a change in the contents.

are you sure your browser-client/servlet-server is doing encoding properly ?

Well, my tables are all UTF-8, but I didn't specify UTF-8 in my DB URL connection (becuase Hibernate was giving me crap about it, I think it had something to do with the &, I didn't really look into it)...

But how can it be encoding issues if everything else is fine?... Here, let me try to type something in Russian...

OUCH!!!!

The Russian is turned into Gibberish, and the page is encoded in UTF-8. Which means that the bridge between Hibernate and DB messes up the encoding...

But that still woudln't explain why English encoding is messed up... Especially in such wierd spots. Although, who knows..

Then the question is - how do I specify &encoding=UTF-8 ?

By the way, I just bought the Hibernate in Action book today (have your donuts, Gaving King and Christian Bauer, hope you don't choke ;)), so I'll look in it as well.

I promise you that hibernate is not messing with any encoding settings or contents AT ALL - its all pure passthrough to the db.

What database are you using?

Are you sure the text you're putting in is UTF-8 encoded?

You promise huh? But do you swear on your mother? :)

In the past I had a bit similar experience (with the exception of random gibberish appearing) where I had the input page in UTF-8, the output page in UTF-8, and the db fields in UTF-8.

BUT! I didn't specify encoding=UTF-8 as a part of the URL, and as a result nothing worked. From that point on I simply added that string as a parameter to the DB URL, and everything worked just fine. When I tried doing the same with Hibernate, i got some error about improperly formatted XML file.. I didn't really look into it though.

But that still wouldn't explain why random crap appears..

Although:

1) THe text is stored in DB in UTF-8

2) Over our non-UTF-8 connection, the client retrieves it, but the page is UTF-8, so it converts some characters in the text field (but leaves ENglish same since it's the same char values in UTF-8 or ascii)

3) Update method is called (every single property is reset beforehand - we're in the "modify object" page)

4) Hibernate sees that what was in database (normal UTF-8) is different from what the user's giving us (UTF-8 filtered through ascii and filtered back into UTF-8), and as a reult it changes it in the database. So, now we have UTF-8 in the database, which has been crippled by ascii..

That's my guess, I guess I'll have to figure out how to append the encoding thing on the DB URL...

Any clues, mates?

I swear yes.

what db are you running on ?

Seriously! Dude! What database are you using?

I'm wondering if you're using MySQL. I've had problems with MySQL and UTF-8 and tracked it down to the driver. It seems to me that the Connector/J 3.1 series is broken where the 3.0 series isn't.

And that thing with the improperly formatted XML file: I bet you added '&encoding=UTF-8'. The ampersand breaks XML because it needs to be terminated with a semi-colon. Try this '&encoding=UTF-8'.

Hey!

I'm on MySQL 4.1

And yeah I'll try the older JConnector and the encoding URL as soon as I finish eating my baby back ribs with El Much Picante sauce. Mmm.

You probably won't need to fiddle with the URL as the connector can determine which encoding to use.

The problem we had was slightly different in that the driver would drop all information after encountering a UTF-8 character.

YUP!

It was the missing UTF-8 part in the URL, not the driver...

Basically, I appended the following to my URL:



And now everything works like a charm!

No gibberish, and all languages supported! (I checked Russian)

Hmm. I might try the same thing then, just to be sure. It's weird because I'd tried that same URL and had no luck with it.

Oh well. Glad you're sorted.