I don't know this is a bug , but in my Turkish forum , convert to latin characters don't work right.It transforms to latin by missing one or two letter.I disabled this feature.Now if I copy links to elsewhere it is look like www.example.com/%C3%BCketiciye-%C3%96zel-%C3%87a%C4%9Fr%C4%B1 . It will be a handicap for seo ? What are your thoughts?
Google still knows to read them. However, URLs, these days, have very minimal to non-existent impact on your SEO. It's all about content these days. I also had non-latin characters in my URLs and it gave me the same structure URLs (I guess the browser converts it to these ugly ones). Anyway, I've disabled it completely and my SEO improved regardless.
Romanization is language agnostic -- what is correct in your language is incorrect in another language. (Ü to "ue" would be correct in German, IIRC.) As such, this is expected. You would need to modify the romanization code to be specific to your language if you wanted to override the behavior.
Ü letters corrent value is u , Ö's is o , Ç'is c . How can I modifiy romanization ? If non romanization letters have no handicap for seo , I can use with like that.But my link look likes example.com/%C5%9F%C3%B6rt/ , if I copy link from adress bar to elsewhere.