====== Bash ======

===== re-run as other user =====

<code bash>
if [ $UID == 0 ]; then
    exec su -c "$0" john
fi
</code>

===== Check if command is installed =====

<code bash>
if ! command -v "${TOOL}" &>/dev/null; then
    echo " Command ${TOOL} not found. Please install it"
    exit 1
fi
</code>

and use it inside loop:
<code bash>
for tool in awk bc sed; do
    ....
done
</code>

===== Output =====

<code bash>
DEBUG() {
    if [ $DEBUG_ENABLE -eq 1 ]; then
    >&2 echo "DBG: $@"
    fi
}
</code>

===== Variables with space =====

<code bash>
# Treat arguments as parts of one argument
declare DST=$*

# Use double colons to pass argument with spaces
nice ionice -c idle btrfs filesystem defragment -v -r -czlib "${DST}"
</code>

Quote problematic characters to use in shell invocation:
<code bash>
  QUOTED_VAR="${VAR@Q}"
</code>


===== Default values =====

<code bash>${parameter:-word}
    If parameter is unset or null, the expansion of word is substituted. 
    Otherwise, the value of parameter is substituted.</code>
    
<code bash>FOO=${VARIABLE:-default}</code>

Or, which will assign to VARIABLE as well:

<code>FOO=${VARIABLE:=default}</code>

===== reading in loop without subshell =====
<code bash>
var=0
while read file
do 
  echo $file; var=1
done < <(ls -1 /tmp/)
</code>

==== 'read' insid loop ====
In loop where stdin is redirected the easiest way to use read is:
<code bash>
TTY=`tty`
while read ...
do
  read CTRLC < ${TTY}
done < somedata
</code>


===== concatenate output (subshell) =====
<code bash>
( command1 ; command2 ; command3 ) | cat
{ command1 ; command2 ; command3 ; } > outfile.txt
</code>

===== leading zeros / octal =====
To interpret a number as decimal, use 10#n form, eg. 10#09

<code bash>
VAR=077; echo $(($VAR+1))
# 64

# specify base method
VAR=077; echo $((10#$VAR+1))
# 78

# strip leading zeros method
VAR=077; VAR=${VAR#0}; echo $(($VAR+1))
# 78
</code>

===== include/source other scripts =====
Possible test scenario:
  - call using from another directory full path /home/user/bin/myscripth.sh
  - add /home/user/bin to PATH and then call from another directory by "myscript.sh"
  - create symlink to script and call using symlink. Check for relative and absolute symlinks.
  - it should work also if it is sourced "source" or "."

The best is to create setup/installer script which hardcode path to one shared script.

Another solutions:
<code bash>
#!/bin/sh
MY_DIR=$(dirname $(readlink -f $0))
$MY_DIR/other_script.sh
</code>

<code bash>
DIR=$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )
</code>

<code bash>
# Works correctly if script is sourced from another one
DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && cd -P "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"
</code>

<code bash>
cur_file="${BASH_SOURCE[${#BASH_SOURCE[@]} - 1]}"
cur_dir="$(dirname "${cur_file}")"
</code>

''`BASH_SOURCE'
     An array variable whose members are the source filenames where the
     corresponding shell function names in the `FUNCNAME' array
     variable are defined.  The shell function `${FUNCNAME[$i]}' is
     defined in the file `${BASH_SOURCE[$i]}' and called from
     `${BASH_SOURCE[$i+1]}'
''



===== double brackets [[ ]] =====
[[http://tldp.org/LDP/abs/html/testconstructs.html#DBLBRACKETS|DBLBRACKETS]]

Bash double brackets [[ ]] (extended test command) are safer and provides more features but they are not portable (and not sh compatible). Features:
  - not filename expansion. Strings will be take literally (i.e. "name*").
  - operators like || instead of -o
  - regex matching =~

<code>[[ -e $file_name ]]</code>
<code>[ -e "$file_name" ]</code>

===== Check if string contains another string =====
not compatible with sh
<code>if [[ $string == *"word1 word2"* ]] ...</code>
<code>if [[ $string =~ .*word.* ]] ...</code>


==== Execute command until fail ====
<code bash>while command; do :; done</code>