Converting UTF8 characters to slash escaped numeric values

Question

My system language is not Japanese, but my user language/encoding is. Using shell things in terminal emulators 'directly' works okay, but it seems that 'behind the scene' stuff does not.

保存 = Save for example.

This gets displayed correctly when I use X stuff 'directly' (as in, I open a terminal emulator or GUI program and mess with it directly) But running stuff in the background (using dmenu, in the xmonad WM ) it turns out that what gets fed into scripts is not something I can string compare to, but something like this:

保存...

gets turned into

'344円277円235円345円255円230円...'

How do I generate this myself? I only got the escaped value string from this by putting a copy paste thing into the script.

I'm fine with doing a pre-pass that translates such an escaped value string into a 'proper' one that then gets case matched later.

I would simply like to generate such an escaped value string myself without the roundabout hack.

• Stéphane Chazelas: Is it just the characters outside the ASCII range you want to convert or all of them. What should happen for control characters (like TAB, NL) or \ or ' themselves?

Answer 1

With the zsh shell, with:

string="保存 = Save for example.
Also including newlines,
'quotes and \backslashes"
() {
 local LC_ALL=C
 quoted=${${(qqqq)1}#'$'}
} "$string"
printf '%s\n' $quoted

Gives:

'344円277円235円345円255円230円 = Save for example.\nAlso including newlines,\n\'quotes and \\backslashes'

If it's only the bytes over 0x80 you want to convert to \ooo:

printf %s "$string" |
 perl -l -0777 -pe 's/[\x80-\xff]/sprintf "\\%o", ord$&/ge'

would give (from any Bourne-like shell):

344円277円235円345円255円230円 = Save for example.
Also including newlines,
'quotes and \backslashes