FUDforum: How To » Message body character set conversion

Home » FUDforum » How To » Message body character set conversion (While upgrading 2.7.7->2.8.1)

Show: Today's Messages :: Polls :: Message Navigator

Message body character set conversion [message #161185]

Sun, 22 November 2009 10:01

nstorm

Messages: 19
Registered: March 2005
Location: Russia, Rostov-on-Don

Karma: 0

Junior Member

Hello.
My forum encoding was koi8-r with 2.7.7 version before. The mysql tables was broken as they had latin1 encoding, but the actual encoding was koi8-r. But it worked up to a 2.7.7 (well, localized search didn't worked, but nobody needed it). I've stopped at 2.7.7 because I was aware what from 2.8.1 it switches to UTF8 by default. But now I finally decided to fix it and update. I wanted to do this step by step, not going straight to 3.0.0. So I've updated to 2.8.1 first. Ended up with broken encoding. First I've fixed template by creating new one with correct locale. Next I've tried MySQL Charset Changer, it changed table encodings but didn't helped as I was expecting, because it was broken with latin1 before. So I've dumped the DB, converted it koi8-r -> utf-8, added set names utf8 and re-imported it. And seems that fixed DB issues, as I can see all categories, topic titles etc right. But message bodies are still in koi8-r. I've read through this forum and tried this from wiki:

The character set conversion feature can be used to change the character set encoding of the message bodies. Non-English forums that are not using UTF-8 are encouraged to convert their messages to UTF-8. Please leave empty if you don't require a character set conversion or if you are unsure. Please note that converting twice, or picking the wrong from and to character sets will corrupt your messages, after which you will have to restore your forum!

First I don't see any option to convert character set when I open message compactor. I assume it works automatically, but running it didn't worked for me. Message bodies still in koi8-r. I could convert them manually, but that will break the indexes, due to various string lengths.
What should I do? And do I run that character set conversion feature for message bodies?

PS: And btw emails from this forum doesn't works now. I've forgot my password and tried to restore it, but it didn't come to my email.

Report message to a moderator

Re: Message body character set conversion [message #161186 is a reply to message #161185]

Sun, 22 November 2009 10:38

naudefj

Messages: 3771
Registered: December 2004

Karma: 28

Senior Member
Administrator
Core Developer

nstorm wrote:

I wanted to do this step by step, not going straight to 3.0.0. So I've updated to 2.8.1 first.

The recommended path is to go from 2.x straight to 3.0.0. You are welcome to first upgrade to 2.8.1, however, free support is only for users on 3.0.0 or that are upgrading to 3.0.0.

nstorm wrote:

First I don't see any option to convert character set when I open message compactor.

You should see it. If not, you're probably not on 2.8.1 or the upgrade failed.

nstorm wrote:

PS: And btw emails from this forum doesn't works now. I've forgot my password and tried to restore it, but it didn't come to my email.

I'm aware of the e-mail problems. The server belongs to Ilia's and I, unfortunately, don't have access to do anything about it.

Report message to a moderator

Re: Message body character set conversion [message #161187 is a reply to message #161186]

Sun, 22 November 2009 10:57

nstorm

Messages: 19
Registered: March 2005
Location: Russia, Rostov-on-Don

Karma: 0

Junior Member

As I said, I'm going to 3.0.0, but want it to be step by step. 2.8.1 first. And the upgrade went fine, my admin panel looks updated, it shows version 2.8.1.
Seems like I've already found where the problem could be by looking at the source code. Seems like to see an option for optional character set conversion, I need to have an iconv php module enabled, which I don't. Will try to enable it now and see.

Report message to a moderator