How to delete a file starting with .#?

[mps2006]

In my home directory I have found a file with a name .#<someletters>.txt.

I have no idea where it came from.

I cannot cat or delete the file - it says "file not found", even though tab compeletion in the shell works!

escaping the characters ie \.\# doesn't work either - it gets automatically re-set to the original name.

Any suggestions? Should I be worried?

thanks
M.

[bonedome]

Hello
can you read it, what are its properties ?

[Dutchy]

What about wrapping the filename in double quotes?

[aleph]

If the filename contain unintelligible or otherwise very strange characters, here's a method to delete them.

First, use the "ls -i" method to list the files and the inode numbers, and copy the offending file's inode num.

After that, use the command "find" to deal with them. For example, if you want to delete it, pass the rm command to the -exec option

find . -inum INODE_NUMBER -exec rm {} \;

where INODE_NUMBER is the one you got from "ls -i". This will descend from the current directory to find the matching file.

This could be very slow, if "find" has to go across filesystem boundaries. To prevent this, add the "-mount" option so that filesystem boundaries are not crossed.

find . -mount -inum INODE_NUMBER -exec rm {} \;

Of course you can use other commands, such as "mv", to rename them.

[Gareth Jones]

Can you see the file in a file manager such as Nautilus or mc, if you enable "show hidden files" (Ctrl+H in Nautilus)? Can you delete it then? A file management program should make the unlink system calls directly, bypassing any shell interpretation etc.

[jpollard]

rm ?rest_of_the_name

[bensonparis]

All above information are right
It is a common methods

[Doug G]

If you use mc from the command line I think it will handle deleting files with weird filenames.

[kurtdriver]

You type rm .#filename.txt using file completion and get an error meassage of "File not found"? Is this correct? You might try temporary move all other .txt file in that directory to another directory and the try again with
rm *.txt.
But you haven't told us if you can display the contents, also if you don't what it is, google the name and see what come up. I did that a few years ago and it turned out to be the conficker worm.

[sea]

Not quite sure, but ":" requires to have an \ previous in order to use it for filenames.
I assume the same works for #...
Asuming that the dot in the file name indicates a hidden file, and should not mark the 'current dir'?

So have a try with: (note the backslash)

rm .\#filename

[jpollard]

One of the problems with all suggestions that use the "#" is that SOMETIMES the character displayed is not the character in the name.

This is the difference between octal 043 (the #) and 243 (also the #).

Also note that # is the start of a comment (so the \ should work, or putting it inside quotes).

One of the harder files to delete is something starting with "-"... quoting these doesn't work as it isn't the shell doing the interpretation. The application identifies it as option specifications.

[aleph]

I think my suggestion using "find" could work with these cases. Human intervention is needed to identify the file, though, but I guess this is inevitable. It is inevitable because you don't know the machine-readable form of the filename in the first place. We have to assume the filename is completely garbage.

[mps2006]

Thanks, those ideas sorted me out.

I'm still a bit worried about where the files came from - there were four, in different directories, both were the names of other innocuous files just with a .# in front of them. one was a .tex and there were .txt files.

They were all sim-links to non-existent files with names of the format user@localhost.13571:1318356082 (so a bit like a maildir message format).

They all had lrwxrwxrwx permissions.

What programme might produce files like that? I'm guessing emacs as that is what I might have used to edit .tex and .txt files.

Thanks
M.

[Gareth Jones]

Originally Posted by mps2006

I'm still a bit worried about where the files came from - there were four, in different directories, both were the names of other innocuous files just with a .# in front of them. one was a .tex and there were .txt files.

They sound like editor back-up/working files. I'm not sure which editor (I don't have Emacs installed to test), GVim uses .filename.swp though.

They all had lrwxrwxrwx permissions.

All symlinks have those permissions (normally) – the real permissions are those of the pointed-to file.

[bigflopper2]

go to the directory and do:

convmv -f iso-8859-15 -t utf8 -r --notest .

all "special characters" will be converted...and you can delete the files...