Bash: shell script to obtain the cipher list of a server

#!/bin/bash

 

SERVER=$1

 

if [[ "$SERVER" == "" ]]; then

  echo "Usage: $0 hostname:port"

  exit

fi

 

DELAY=2

ciphers=$(openssl ciphers 'ALL:eNULL' | sed -e 's/:/ /g')

 

for cipher in ${ciphers[@]}

do

  echo -n Testing $cipher ...

  result=$(echo -n | openssl s_client -cipher "$cipher" -connect $SERVER 2>&1)

 

  if [[ "$result" =~ "Session-ID:" ]]; then

    echo "YES"

  else

    if [[ "$result" =~ ":error:" ]]; then

      error=$(echo -n $result | cut -d':' -f6)

      echo "NO \($error\)"

    elif [[ "$result" =~ "errno=104" ]]; then

      echo "NO \(Connection reset by peer\)"

    else

      echo "Unknown response"

    fi

  fi

 

  sleep $DELAY

done

 

Backreference in Perl vs sed

Perl uses $n (where n is a number) for backreferences; while sed uses \n.

For example, we want to switch the contents between #:# in this file:
#Hello:World#!
How #are:you#?

Using sed, we run command:
$ sed -n -e 's/#\([^:]*\):\([^#]*\)#/#\2:\1#/p' filename

Using perl, we can run command:
$ perl -p -e 's/#([^:]*):([^#]*)#/#$2:$1#/' filename

or:
$ perl -p -e 's/#([^:]*):([^#]*)#/#${2}:${1}#/' filename

The output will be:
#World:Hello#!
How #you:are#?

A few notes:
1. The backreference of sed can be only one digit. That means \10 will be \1 (the first reference) and a digit 0.
2. sed uses \( and \) to group while Perl uses ( and ) only.

Linux: find and replace text across multiple lines

For example, we have a file myfile.txt with the following content:

This is
my
apple
in
the basket.

We can run this command:

$ perl -0777 -i -pe 's/This is\nmy\napple/These are\nmy\napples/g' myfile.txt

to change the content of the file into:

These are
my
apples
in
the basket.

The option -0777 applies the matching to the whole file instead of line by line. The option -i makes the changes in place.

Linux: trim the extensions from the filenames for a batch of files

Trim the extension (e.g. .bak) of one file is simple, e.g.:

$ mv myfile01.txt.bak myfile01.txt

If we have hundreds of such files, nobody would want to run hundreds of the similar commands:

$ mv myfile02.txt.bak myfile02.txt
$ mv myfile03.txt.bak myfile03.txt
...

We can use sed instead:

$ ls *.bak | sed -e 's/\(.*\)\(\.bak\)/mv \1\2 \1/' | sh

The explanation of this command:

1. "ls *.bak" gets the list of the filenames and feed them to sed.
2. For each filename, sed changes it to "mv something.bak something"
3. In the first part of the "s" command of sed, a pattern matches anything ended with .bak. It uses \( and \) to group the filename into 2 parts. The 1st group \(.*\) matches anything before .bak. The 2st group \(\.bak\) matches the exact string .bak.
4. In the second part of the "s" command of sed (mv \1\2 \1), \1 and \2 are substituted with the 2 groups described above. So \1 refers to anything before .bak, and \2 refers to .bak.
5. The output of sed will be "mv something.bak something" and it is passed to sh command to run.
6. This will be done for all the files from 'ls *.bak'.