This bug is fixed for 1.4-DR8.

Ingomar,

We'll attack this for 1.4 - I think we should be able to fix it now Charles' has tracked down the problem in detail. Sorry for taking so long, it's one of those things us non-i18n people don't think of (that the lowercase version of a character might be different in Java and in Postgres!!!).

m

I've found the problem. We do a case-insensitive match to find pages by their title: so Bob and BOB and bob are the same page.

The problem is that we do this by calling title.toLowerCase() in Java code, but matching against lower(title) in the database. So if Java and the database disagree on what the lower-case of a particular letter is, then the whole thing blows up. Viz:

confchars=# select lower('ÄÜÖ');
lower
-------
ÄÜÖ
(1 row)

We should change the query so that we do the toLowerCase() in the query itself:

select * from content where lower(title) = lower('ÄÜÖ');

Mhhm, I am not sure whether I am about to loose my sense of humor about this.

Ingomar,
Yes, fair point; i18n is currently scheduled for 1.5.

Cheers,
Nick

–
Nick Faiz - nick@atlassian.com

ATLASSIAN
Confluence - the professional J2EE wiki - tried it yet?
http://www.atlassian.com/confluence

JIRA - need a simple, powerful way to track and manage issues?
http://www.atlassian.com/software/jira

This issue now celebrates it's 0,5 bithday.

The problem persists in 1.3.
I don't think this has anything to do with the DB, as it works fine for lowercase charcters. I would assume that this is an encoding issue.

I have another observation:

Same scenario, I jsut created that page.
The i go back to the dashboard.
The page shows up on "recently update" - when I select THAT link everything is fine.
When I go to the page I created that page from (by creating a link and selecting that one) I get the silly "create page".

I could reproduce this in the public test space:
http://confluence.atlassian.com/display/TEST/CONF-1570+Test

I am sorry to report that although I did the migration as suggested I still recieve the same error.

Here is the database i use right now:
cfu | cfu | UNICODE

Ah, this is most likely the problem. You have created your databases in "unknown 8 bit character set" mode, so they're behaving randomly when presented with multi-byte Unicode characters.

Your should probably do the following:

Back up your Confluence database (using pg_dump). Drop the database and then re-create it with Unicode support using 'createdb -E UNICODE'. Import your dump-file into the new database.

I'm not entirely sure what effect this will have on existing high-bit or multi-byte data in the database, though.