(not really) easy

[iribach]

hi all,
is there an EASY and SURE way to translate (stupid)filenames containing spaces, [ ` ^ $ \ ; &quot; ' & * % | > < all-brakets ] + special chars >127 and so on
to standard unix filenames, changing this characters BUT preserving the (unique)name ? exemple:

'\\ab` \234 ab' -> __ab1____x_ab
'\\ab^ \234 ab' -> __ab2____x_ab

sure in c, using opendir(), readdir(), char *xx on filename to check all characters and changing them, then re-readdir() to assure the new name
is unique ... i am looking for an one-liner sh|sed|awk|perl script.
thanks a lot.
ira
PS: we have a lot of users not knowing, they are writing dos-files on a unix server

 

[dickiebird]

Look at man pages for tr
or try this ( untested ! )

for file in * # you decide on the range
do
mv $file `echo $file | tr -c '[A-Z][a-z][0-9]' '[_*]' `
done

HTH
Dickie Bird (-)))

 

[iribach]

thanks dickiebird,
tr is a good utility, but what about unique filenames ?
'\\ab` \234 ab' -> __ab1____x_ab
'\\ab^ \234 ab' -> __ab2____x_ab

 

[dickiebird]

I would try tr twice piped :

for file in a????
do
mv $file `echo $file |tr '\`\^\`\-\_' '12345'| tr -c '[A-Z][a-z][0-9]' '[_*]' `
done

i.e Replace up to 10 punctuations with 0 to 9 then tr the complement of that.
NB Still untested

HTH Dickie Bird (-)))