Question

TWiki 4.1.1 on FreeBSD/apache1.x/perl5.8.8, no extra plugins etc, new install, supporting Chinese, Pinyin, English.

Set up TWiki following all relevant docs for UTF-8 and heeding warnings about not setting charset. Only need ASCII English wiki words, but multiple languages in text.

Chinese and pinyin work fine, save and re-edit works, but the editing is hell.

The problem is that in the edit screen, all of the Chinese and half of the pinyin is html entities.

• a section of view topic:
  

• part of the edit screen:
  

... ... ... ... ...

Using the default wysiwyg editor, it looks fine during edit, but I don't want to have to use that one.

I know it can work, because I tried entering and editing the same data on a Chinese UTF8 TWiki and it worked well, the simple editor's contents looked the same as the displayed topic. (I worked in two tabs of the same browser session, pasting the same clipboard data into each TWiki.)There is no info to contact the admin there, and I don't know much useful Chinese to ask about their setup.

Clearly I'm missing some little thing that makes this work properly on TWiki, but what is it?

Environment

-- SueBlake - 20 Feb 2007

Additional info

Now I'm thinking that all I need is to get UTF-8 sent on every view and edit.

I've tried with UseLocale off, and with it on and set to UTF-8, and set to en_us.ISO8859-1, charset untouched (defaults to iso-8859-15).

I've put "AddDefaultCharset UTF-8" into httpd.conf and tried again with and without UseLocale, also tried with utf-8 (lower case)

I've put "AddDefaultCharset UTF-8" into .htaccess and tried again with and without UseLocale, also tried with utf-8 (lower case)

All no use. The headers still won't send UTF-8 but stick with sending iso-8859-15. Tested with both Firefox and Lynx.

So where's the crowbar that forces UTF-8?

Answer

If you answer a question - or have a question you asked answered by someone - please remember to edit the page and set the status to answered. The status is in a drop-down list below the edit box.

The behaviour of TWiki 4.1.2 is not as documented, you do need to type in the Site Char Set at the moment, otherwise you get the results above. See Bugs Item3751

I've put UTF-8 into configure and it's working much better, no more numbers. I have some other UTF-8 problems now but I'm closing this one.

-- SueBlake - 13 Mar 2007