Hi, In an old text file, which "file" says it is utf-8 text, there are some chars which are not. Some were obviously accented letters, like "á", so I just did a search and replace on them. There is one entry I don't know what it is: ············ rsync -a myfolder/*/*.jpg my-new-folder/ ············ (The file got corrupted at some point because one editor thought the file was utf-8, and another editor thought differently). I saved the bad part to a file, but the command "file" insists it is utf-8 cer@Telcontar:~> cat p ············ ············ cer@Telcontar:~> file p p: UTF-8 Unicode text cer@Telcontar:~> And I fail to force a conversion: cer@Telcontar:~> iconv -f LATIN6 -t UTF-8 p ÷÷÷÷÷÷÷÷÷÷÷÷ ÷÷÷÷÷÷÷÷÷÷÷÷ cer@Telcontar:~> Perhaps I'm guessing wrong the non utf encoding. I don't remember for sure which was the old latin encoding we used in Spain, too. My guess is that the string is ················ which is a centered dot, which in my keyboard (Spain) is on the [.] key, but pressing also [AltGr] Testing besides the bad string: ··· Ideas? -- Cheers / Saludos, Carlos E. R. (from Telcontar, using openSUSE Leap 15.4)