The following is a snippet of a larger script I'm attempting. I just want this part to recognize the argument is a directory and then cd to that directory: i.e ./larj /etc.
#!/bin/ksh
# Filename: larj.sh
if [ $# -gt 1 ]; then
echo "0 or 1 arguments allowed."
exit
fi
if [ -f "$1" ]; then
echo "directory only."
exit
else
if [ -d "$1" ]; then
cd $1
fi
fi
When I run the script with /etc as the argument, it appears nothing happens; it stays in the same directory with no error.
Anyone have any suggestions how to get it to change directories?
Thanks
The cd is taking place within the script's shell.
When the script ends, it's shell exits, and you return to the directory before running the script. In order to change the directory you can
mkdir testdir
. ./your_script.sh testdir
At the end of the script you will be moved at directory testdir.
The problem why you cd can't work is that cd executes in the sub-shell when you execute the script as ./larj /etc. So when you execute the script, it changes the working directory of the subshell and has no impact on the current shell.
So you can execute it as . ./larj /etc.
Refer to Why doesn't “cd” work in a bash shell script?.
Related
I have the problem that my shell script is not acting exactly the same as my manual typing into a console. I am attempting to find and source some setup files in a shell script as follows:
#!/bin/bash
TURTLE_SHELL=bash
# source setup.sh from same directory as this file
_TURTLE_SETUP_DIR=$(builtin cd "`dirname "${BASH_SOURCE[0]}"`" > /dev/null && pwd)
. "$_TURTLE_SETUP_DIR/turtle_setup.sh"
This bash file calls a .sh file:
#!/bin/env sh
_TURTLE_ROS_SETUP_DIR=$_TURTLE_SETUP_DIR/../devel
if [ -z "$TURTLE_SHELL" ]; then
TURTLE_SHELL=sh
fi
if [ -d "$PX4_FIRMWARE_DIR/integrationtests" ]; then
if [ -f "$PX4_FIRMWARE_DIR/integrationtests/setup_gazebo_ros.bash" ]; then
. "$PX4_FIRMWARE_DIR/integrationtests/setup_gazebo_ros.bash" "$PX4_FIRMWARE_DIR"
fi
fi
if [ "$TURTLE_SHELL" = "bash" ]; then
if [ -f "$_TURTLE_ROS_SETUP_DIR/setup.bash" ]; then
source $_TURTLE_ROS_SETUP_DIR/setup.bash
fi
else
if [ "$TURTLE_SHELL" = "sh" ]; then
if [ -f "$_TURTLE_ROS_SETUP_DIR/setup.sh" ]; then
source $_TURTLE_ROS_SETUP_DIR/setup.sh
fi
fi
fi
The line in question is:
. "$PX4_FIRMWARE_DIR/integrationtests/setup_gazebo_ros.bash" "$PX4_FIRMWARE_DIR"
I have made sure that this code is actually running and that my environment variables are correct. If I run this command on the command line everything works well. However, the same is not true when the file is sourced via shell script. Why is this? Is there something different about the environment of a shell script that is different from a command line. Also, how can I fix this problem?
Edit:
I am sourcing either the .bash or the .sh scale, depending upon which shell I am using.
Edit 2:
I am sourcing this script. Thus, everything is run in my default bash terminal, and it is all run within the same terminal and not a terminal spawned from a child process. Why is the script not sourcing setup_gazebo_ros.bash within the current shell?
It's the same reason why you source the env script and not run it. When you run the script it runs in a new shell and the variables are not transferred back to the parent shell.
To illustrate
$ cat << ! > foo.sh
> export foo='FOO'
> !
$ chmod +x foo.sh
$ ./foo.sh
$ echo $foo
$ source ./foo.sh
$ echo $foo
FOO
I'm trying to create a bash script to setup my development environment. The script is running as root but I get the error line 11: ln: command not found
#!/bin/bash
#Require script to run as root - doesn't work - syntax error in conditional expression: unexpected token `;'
#if [[ $(/usr/bin/id -u) -ne 0]]; then
# echo "Script must be run as root";
# exit;
#fi
#PHPMyAdmin
PATH="/etc/apache2/sites-available/phpmyadmin.local";
if [ ! -a PATH ]; then
ln -s /home/user/Ubuntu\ One/htdocs/vhosts/phpmyadmin.local PATH;
a2ensite phpmyadmin.local;
fi
PATH=...
Congratulations, you've clobbered how the shell finds commands. Don't do that.
PATH tells the shell where to look for commands. In your case, it looks for ln somewhere in /etc and predictably doesn't find it there.
You should use a different name.
I am writing my first Bash script and am running into a syntax issue with a function call.
Specifically, I want to invoke my script like so:
sh myscript.sh -d=<abc>
Where <abc> is the name of a specific directory inside of a fixed parent directory (~/app/dropzone). If the child <abc> directory doesn't exist, I want the script to create it before going to that directory. If the user doesn't invoke the script with a -d argument at all, I want the script to exist with a simple usage message. Here's my best attempt at the script so far:
#!/bin/bash
dropzone="~/app/dropzone"
# If the directory the script user specified exists, overwrite dropzone value with full path
# to directory. If the directory doesn't exist, first create it. If user failed to specify
# -d=<someDirName>, exit the script with a usage statement.
validate_args() {
args=$(getopt d: "$*")
set -- $args
dir=$2
if [ "$dir" ]
then
if [ ! -d "${dropzone}/targets/$dir" ]
then
mkdir ${dropzone}/targets/$dir
fi
dropzone=${dropzone}/targets/$dir
else
usage
fi
}
usage() {
echo "Usage: $0" >&2
exit 1
}
# Validate script arguments.
validate_args $1
# Go to the dropzone directory.
cd dropzone
echo "Arrived at dropzone $dropzone."
# The script will now do other stuff, now that we're in the "dropzone".
# ...etc.
When I try running this I get the following error:
myUser#myMachine:~/app/scripts$ sh myscript.sh -dyoyo
mkdir: cannot create directory `/home/myUser/app/dropzone/targets/yoyo': No such file or directory
myscript.sh: 33: cd: can't cd to dropzone
Arrived at dropzone /home/myUser/app/dropzone/targets/yoyo.
Where am I going wrong, and is my general approach even correct? Thanks in advance!
Move the function definitions to the top of the script (below the hash-bang). bash is objecting to the undefined (at that point) call to validate_args. usage definition should precede the definition of validate_args.
There also should be spacing in the if tests "[ " and " ]".
if [ -d "$dropzone/targets/$1" ]
The getopt test for option d should be-:
if [ "$(getopt d "$1")" ]
Here is a version of validate_args that works for me.
I also had to change the drop zone as on my shell ~ wouldn't expand in mkdir command.
dropzone="/home/suspectus/app/dropzone"
validate_args() {
args=$(getopt d: "$*")
set -- $args
dir=$2
if [ "$dir" ]
then
if [ ! -d "${dropzone}/targets/$dir" ]
then
mkdir ${dropzone}/targets/$dir
fi
dropzone=${dropzone}/targets/$dir
else
usage
fi
}
To pass in all args use $* as parameter -:
validate_args $*
And finally call the script like this for getopt to parse correctly-:
myscript.sh -d dir_name
When invoked, a function is indistinguishable from a command — so you don't use parentheses:
validate_args($1) # Wrong
validate_args $1 # Right
Additionally, as suspectus points out in his answer, functions must be defined before they are invoked. You can see this with the script:
usage
usage()
{
echo "Usage: $0" >&2
exit 1
}
which will report usage: command not found assuming you don't have a command or function called usage available. Place the invocation after the function definition and it will work fine.
Your chosen interface is not the standard Unix calling convention for commands. You'd normally use:
dropzone -d subdir
rather than
dropzone -d=subdir
However, we can handle your chosen interface (but not using getopts, the built-in command interpreter, and maybe not using GNU getopt either, and certainly not using getopt as you tried to do so). Here's workable code supporting -d=subdir:
#!/bin/bash
dropzone="$HOME/app/dropzone/targets"
validate_args()
{
case "$1" in
(-d=*) dropzone="$dropzone/${1#-d=}"; mkdir -p $dropzone;;
(*) usage;;
esac
}
usage()
{
echo "Usage: $0 -d=dropzone" >&2
exit 1
}
# Validate script arguments.
validate_args $1
# Go to the dropzone directory.
cd $dropzone || exit 1
echo "Arrived at dropzone $dropzone."
# The script will now do other stuff, now that we're in the "dropzone".
# ...etc.
Note the cautious approach with the cd $dropzone || exit 1; if the cd fails, you definitely do not want to continue in the wrong directory.
Using the getopts built-in command interpreter:
#!/bin/bash
dropzone="$HOME/app/dropzone/targets"
usage()
{
echo "Usage: $0 -d dropzone" >&2
exit 1
}
while getopts d: opt
do
case "$opt" in
(d) dropzone="$dropzone/$OPTARG"; mkdir -p $dropzone;;
(*) usage;;
esac
done
shift $(($OPTIND - 1))
# Go to the dropzone directory.
cd $dropzone || exit 1
echo "Arrived at dropzone $dropzone."
# The script will now do other stuff, now that we're in the "dropzone".
# ...etc.
I have an array variable tht contains all directories in a folder.
I need to cd to each directory in the $array. bt when it reach to a non accessible directory , program halts . how can I avoid this so that my loop simply go to the next directory. code is
foreach dir ($array)
cd $dir
echo "directory is $dir"
cd - end
please help !!
Just test that the cd worked.
( if cd $dir 2> /dev/null; then
echo in directory $dir
# Do other things
fi )
Also, note the parentheses. These cause the entire clause to run in a subshell, so there's no need to cd back to the original location.
If iam not mistaken you can use something like
if [ -r ${dir} ]
then
echo "Have read access!"
fi
You can use -w for testing write access and -x for testing for the exceution bit.
I'm creating a bash script that will run a process in the background, which creates a socket file. The socket file then needs to be chmod'd. The problem I'm having is that the socket file isn't being created before trying to chmod the file.
Example source:
#!/bin/bash
# first create folder that will hold socket file
mkdir /tmp/myproc
# now run process in background that generates the socket file
node ../main.js &
# finally chmod the thing
chmod /tmp/myproc/*.sock
How do I delay the execution of the chmod until after the socket file has been created?
The easiest way I know to do this is to busywait for the file to appear. Conveniently, ls returns non-zero when the file it is asked to list doesn't exist; so just loop on ls until it returns 0, and when it does you know you have at least one *.sock file to chmod.
#!/bin/sh
echo -n "Waiting for socket to open.."
( while [ ! $(ls /tmp/myproc/*.sock) ]; do
echo -n "."
sleep 2
done ) 2> /dev/null
echo ". Found"
If this is something you need to do more than once wrap it in a function, but otherwise as is should do what you need.
EDIT:
As pointed out in the comments, using ls like this is inferior to -e in the test, so the rewritten script below is to be preferred. (I have also corrected the shell invocation, as -n is not supported on all platforms in sh emulation mode.)
#!/bin/bash
echo -n "Waiting for socket to open.."
while [ ! -e /tmp/myproc/*.sock ]; do
echo -n "."
sleep 2
done
echo ". Found"
Test to see if the file exists before proceeding:
while [[ ! -e filename ]]
do
sleep 1
done
If you set your umask (try umask 0) you may not have to chmod at all. If you still don't get the right permissions check if node has options to change that.