Wrong character encoding after editing

[RubioTerra]

This problem was reported on the Portuguese section of the forum and I couldn't answer it. The user wants to edit a dictionary module, adding info to existing topics. But whenever he adds any text containing diacritics (and Portuguese has quite a few), although the text looks OK at first, all accented characters becomes garbled once he leaves the topic and goes back.

I attach here a sample of the module. To reproduce the problem, just add any non-ascii text to a topic, switch to another topic and then go back. All non-ascii text, including text that existed before the editing, becomes wrong. I checked the RTF, it has no font tables and all non-ascii characters are encoded as \uXXXX?.

[JG]

It appears that the sample database date is in UTF-16le encoding. I am not sure if Sqlite is supposed to deal with this or if it is theWord. A workaround to the issue appears to be to drag and drop the topics to a new user module within theWord. However I would wait for Costas more informed reply.

[RubioTerra]

Thanks, JG. This workaround does away with the problem. I'll copy all topics to a new module and send it to the user that reported the problem.

[csterg]

theWord will only deal with utf8 databases. This is a system error with the db itself, i don't plan to change it
Costas

[RubioTerra]

That's alright. It's rare to find UTF-16 content anyway.