Audacity Forum

Gale Andrews wrote: Once again, that isn't what I found when I tested this recently with German and East Asian characters on English Windows

I was referring to the specific characters used in this example.

The AUP file had the character "è" but the file is ANSI encoded, and this caused an "Invalid Token" error in both the Unicode version of Audacity and the ANSI version. However, if I copy and paste the track name ("Zvoèna sled") as a track name into a new project with Audacity 1.3.14 Unicode and Save the project, then it opens correctly in both Unicode and ANSI versions of Audacity (though the track name is displayed as "Zvoèna sled")

If I open the original AUP file in a text editor and change the page encoding to UTF-8, then the Invalid Token error does not occur in either version of Audacity.
If I then save that file with the ANSI version of Audacity and then try to reopen it, the Invalid Token error still does not occur with either versions of Audacity, but the track name has changed to "Zvoèna sled".

I don't think that my observations contradict Gales observations and I agree that a critical question is how those characters got into the project in the first place.
Also, the advice remains the same - avoid accented characters (in both track names and Metadata) when using an ANSI build of Audacity. If it is necessary to use accented characters then stick with the Unicode build.

steve wrote:I was referring to the specific characters used in this example.

Everything you mentioned in your post above accords with my tests too. It was just that you said if you save Unicode characters in ANSI Audacity and open the project in Unicode Audacity "the problem will still occur". I read that as the invalid token error occurring (the subject of the topic). All we can reproduce in that case is the characters being displayed incorrectly in the program.

I guess it may need someone actually using a localised version of their OS and a localised keyboard to reproduce the error. It is reproducible by pasting "è" directly into an ANSI encoded .aup file, but I doubt many people are doing that.



Gale

Gale Andrews wrote: It is reproducible by pasting "è" directly into an ANSI encoded .aup file, but I doubt many people are doing that.

I agree. The question is as you said (asked) before - how did they get the invalid characters into the project? I don't think that we have an answer to that yet, do we? I've tried on Win XP with Audacity 1.3.13 ANSI and I've not been able to make an AUP with invalid characters as they always seem to be translated into valid equivalents.