Character sets

[hans-peter.geier]

While I try to convert a twiki into Confluence (on Linux Redhat) using UWC (version 3.12), I'm facing the issue that UWC is not handling foreign characters correctly, e.g. German Umlaute, French accents and such.

The twiki file has its special characters encoded to iso88591, and I have set the $LANG of the Linux shell environment to en_US.iso88591 accordingly.

The local shell tools (like vi) display the special characters correctly.

The UWC output, though, (both, the local output files as well as the uploaded files in Confluence) show these special characters unreadable. (even in vi)

(I have also attempted to set it to de_DE.iso88591 without getting any change in results)

How to make uwc to consider the language-specific characters? Is there anything I can do by exporting the appropritate environment variable (or similiar), or is UWC simply not supporting these characters?

[Laura Kolker · Accepted solution]

The UWC is definitely capable of handling ISO-8859-1. I've tested it against the common encodings that don't overlap nicely with UTF-8, including ISO-8859-1, ISO-8859-15and Cp1252. To start, you should make certain you have set the encoding property in your converter.twiki.properties file. Here's some doc on that:

UWC Character Encoding Feature

That will probably handle the problem, but here are some other ideas, if the above is insufficient:

• check your confluence encoding setting in your admin console, also your confluence database settings
• you can also use a tool like iconv to transform the encoding of your pages into UTF-8 before using the UWC on them. Some examples on that idea can be found in the UWC FAQ on encoding

Good luck!

Laura

[hans-peter.geier]

Thank you! I got it to work.

[Christian Janzen]

Hi Laura,

I tried to convert a SharePoint wiki and have the same "german umlaut problem". But entering Sharepoint.1234.encoding.property=iso-8859-1 at the end of the file converter.sharepoint.properties doesn't work. Even Sharepoint.1234-encoding.property=iso-8859-1 is not working.

Any other idea? I am working with version 3.13.0

Best regards,

Christian