Question

I am trying to migrate data from Windows to Ubuntu on a non-english speaking basis. Therefore TWiki is configured on both systems following the manual on InstallationWithI18N. Independently from existing data, WikiWord-linking on each system works fine.

Observation: When a new topic generated on Ubuntu is created and the topic name includes characters like ó, e.g. PruebaTraducción, the topic name and corresponding links look good, but the corresponding filename in data looks like PruebaTraducci?n.txt.

So here comes the problem:
When I try to migrate the data from Windows to Ubuntu, the content shows up fine, but all WikiWords and corresponding links do not work. I assume that it has to do with the strange filenames, too, but I have no idea how to straight this out. Any help is really appreciated - thanks in advance.

Environment

-- SebastianKlus - 29 May 2008

Answer

If you answer a question - or someone answered one of your questions - please remember to edit the page and set the status to answered. The status selector is below the edit box.

Are the Windows and Ubuntu implementations both on TWiki 4.2?

If the Windows one is older, that could explain why the problem only shows up on Ubuntu. KnownIssuesOfTWiki04x02 includes TWikibug:Item4946, which refers to I18N problems.

-- SeanCMorgan - 29 May 2008

No, both are TWiki 4.2. I suppose it has to do with character encoding. But I am not sure.

-- SebastianKlus - 30 May 2008

I partly figured it out, although I am not sure if either I did something wrong during the configuration of I18N or it shall just work this strange way:

I had to convert all filenames from uft8 (they got there like this when I copied them from Windows via USB-stick to Ubuntu) to ISO-8859-1 using convmv. It worked fine - now all filenames with characters like á look strange But TWiki works with all its links and WikiWords.

Another small details that attracked my attention: I have a drop down box in the WebTopBar that allows to change the language from English to Spanish and vice versa. When I am in the TWiki Web, the entrys of the drop-down list show up as English and Español. Now, when I change to one of the custom webs copied from the Windows system, they show up as English and Español.

So I suppose everything is circling around the configuration of I18N and the conversion TWiki does sometimes from one character encoding to the other.
I will keep the topic marked as AskedQuestion to see if anyone has any additional observation on this topic, but if not, I will switch it to answered.

-- SebastianKlus - 31 May 2008

Setback
Normally I edit everything in Raw View, but by chance I used the WYSIWYG editor today and there all accented words look like e.g. b�sicos whereas in Raw View the text looks good (básicos) and modifications are no problem. Any ideas?

-- SebastianKlus - 31 May 2008

Update your WysiwygPlugin and TinyMCEPlugin to the latest, and make sure your Locale and Site Charset are set up correctly in configure.

-- CrawfordCurrie - 31 May 2008

That was it. Thanks.

I still would like to know if it is "normal" to have to convert the filenames from UTF to ISO, but I suppose it is, due to the fact that ISO is set in configure and UTF is not yet an option according to the various manuals.

Just for curiosity (and a little bit off topic): Why can the files inside the data/TWiki folder be interpreted correctly although they are in UTF? When and how does TWiki "decide" to pick a UTF-encoded file and convert it to the configured encoding and when does it have problems with that?

-- SebastianKlus - 03 Jun 2008