Cygwin and Bash????

Beppe83 · 06-22-2004, 03:02 AM

I did a bash script with Cygwin under Windows.

When I open it in Linux Mandrake and I type:

./file.sh

It says:

bad file or interpreter : no such file or directory

So I create a new file .sh from an editor inside Linux, everything goes ok...

I'm confused...

Cerbere · 06-22-2004, 05:16 AM

Probably those pesky DOS carriage return / newlines. Try the following script (from the Advanced Bash-Scripting Guide):

Code:

Example 12-19. Du: DOS to Unix text file conversion.

   1 #!/bin/bash
   2 # Du.sh: DOS to UNIX text file converter.
   3 
   4 E_WRONGARGS=65
   5 
   6 if [ -z "$1" ]
   7 then
   8   echo "Usage: `basename $0` filename-to-convert"
   9   exit $E_WRONGARGS
  10 fi
  11 
  12 NEWFILENAME=$1.unx
  13 
  14 CR='\015'  # Carriage return.
  15            # 015 is octal ASCII code for CR.
  16            # Lines in a DOS text file end in CR-LF.
  17            # Lines in a UNIX text file end in LF only.
  18 
  19 tr -d $CR < $1 > $NEWFILENAME
  20 # Delete CR's and write to new file.
  21 
  22 echo "Original DOS text file is \"$1\"."
  23 echo "Converted UNIX text file is \"$NEWFILENAME\"."
  24 
  25 exit 0
  26 
  27 # Exercise:
  28 # --------
  29 # Change the above script to convert from UNIX to DOS.

You're going to have to remove the line numbers yourself (I'm too lazy), and by the way, the exercise at the bottom is optional ;-)

Enjoy!
--- Cerbere

Beppe83 · 06-22-2004, 06:39 AM

THANKS!

This helped me! Now the script works under Linux too!!!

THANKS AGAIN MAN!

Cerbere · 06-23-2004, 02:24 AM

My pleasure. You might also do a google search for 'dos2unix'. It'll find a bunch of programs and a really elegant perl script for doing the same thing.

Enjoy!
--- Cerbere

cck23 · 06-23-2004, 03:54 AM

Hi

I don't know if this helps but I have always converted DOS files that have the pesky ^M character using the following:

sed 's/^M//' olddosfile > newfilename

(Note: ^M = CTRL V, then M. Dont let go of CTRL)

This seems to me a lot easier than above, but maybe there is something that the script would do that the above sed script wouldn't do? I'd like to know if that is the case

On another note, Beppe83, did you fix your "sort" problem that you posted elsewhere?

CK

wirawan0 · 07-12-2004, 11:01 AM

Or, even simpler:

tr -d '\r' < file.dos > file.unix

for file.dos being the DOS text file to be converted. This will strip ALL carriage returns, which means that at rare occassion it isn't desirable. The best thing would be:

sed -e 's|\r$||' < file.dos > file.unix

These should be run under bash. Other shells may or may not work. Always check first with a test file.