Hi all,

I am running a linux server that is used within my intranet (samba and nfs) as well as from the internet (sftp/ssh).
Some users use their sftp clients with wrong filename encoding what leads to ugly filenames (and problems) for the intranet users.

I don't see any chance to avoid this problem, so I try to fix it automatically. And ran into the next problem: I don't figure out, how to detect files with bad filenames.

I recognized, that finds all filenames, including those with bad filenames, while only finds files with good filenames. So I could compare these two filelists and hope to catch all bad filenames.
No solution I like. First: I don't really know, why find behaves this way, and in consequence, I don't know, if this procedure is very reliable, and second: It seems to generate much overhead, I think.

Does anyone know, how to get "bad" filenames?

Thanks a lot,
Clemente

You could use a bit of grep's regular expression handling for this in combination with your 'find .', or actually find's regexp itself, if you're comfortable.

If I'm not mistaken, \w will only match valid word characters. Thus, if you search for files that do not match [\w\d\-\_\.], you'll have a fairly complete list of files that don't work. If you have a lot of files with other characters in the name (#, @, $, et cetera), you can adjust your regular expression accordingly.

Hope that is a step in the right direction for you!

Thank you very much, I will try this!

Have you tried ls -b? This will show "strange" chars in octal notation.

You almost answered this yourself.
If...
find .
gives everything and
find . -iname '*'
gives good names .... then by extension trying
find . ! -iname '*'
does as you wish... it return files with bad names.
But thanks, you showed me something more about the quirks of find that I didn't know.