Need a little search script to find ^L and ^M

[Weary in West Chester]

Hi Folks,

I need a little help again. Can anyone suggest a way to run through a directory and search the files to see if they have any carriage return / line feeds in them. We get them in some of our text files when when transfer them in binary mode from our unix boxes to our LAN. The dos2unix utility can strip them off if I can identify them 1st. Any suggestions? Thank you in advance.

[Cheryl Griffin]

# ll | cat -vet

Cheryl

[Cheryl Griffin]

Here's a test:
I created a file with vi:
# vi ^l^y^t (by hitting the control L, etc.)
this is a test
:wq!

# ll <--- does not show the file
# ll | cat -vet
total 1124$
-rw-rw-rw- 1 root sys 16 Oct 17 14:52 ^L^Y^T$ <-- here is my file

Here's what a "regular" file looks like:
-rw------- 1 root sys 297 May 13 09:18 .TTauthority$

The $ marks the end of the file name.
The cat -vet helps print those unseen characters.

Cheryl

[A. Clay Stephenson · Accepted solution]

Here's one method off the top of my head:

find . -type f -print | while read FNAME
do
awk '/.*\015/ { exit 99 }; /.*\014/ { exit 98 }' ${FNAME}
STAT=${?}
if [[ ${STAT} -eq 99 || ${STAT} -eq 98 ]]
then
echo "File ${FNAME} needs some work."
fi
done

[John Poff]

Hi,

If I understand your question, you just want to find the files that have carriage return characters in them and run the dos2ux utility on them. Here is a little script I came up with to do it. To get the ^M character in the grep line, I had to hit Ctrl-X and then Ctrl-M. I also had to do it in sh as I couldn't get a ^M entered in ksh, but that is another problem. Take a look at this:


#!/bin/sh

# find_dos_files

for f in $(ls *.txt)
do
CR=$(grep -c '^M' $f)
if [[ CR -gt 0 ]]
then
echo "The $f file has CR/LF and needs to be converted."
mv $f $f.tmp
dos2ux $f.tmp >$f
rm $f.tmp
fi
done




JP

[H.Merijn Brand (procura]

*untested*
: find all files down from the current folder that contain any characters from ^A to ^
# find . -type d | perl -nle'opendir D,$_;for$f(grep/[\001-\037],readdir D){$f=~s/([\001-\037])/sprintf"^%03o",ord$1/ge;print "$_/$f"}'

[Steve Lewis]

To list the filenames containing the sequence under /dir, this worked for me:

echo '^M' > x
find /dir -type f | xargs grep -lf x {}

To get the '^M' use ctrl-V ctrl-M

[Weary in West Chester]

Thank you everyone. I found my problem files and got them fixed. Good job!