We Keep Coding

Using Python to run a sourced script, cd, and then run a command within the SAME shell

How can run a sourced bash script, and then change directories, and then run a command, all within the same shell (Using python)? Is this even possible?
My Attempt:
subprocess.check_call(["env -i bash -c 'source ./init-build ARG'", "cd ../myDir", "bitbake myBoard"], shell =True)

I would make this for you, but I need to see the absolute paths. Here is an example
subprocess.check_call(["""/usr/bin/env bash -c "cd /home/x/y/tools && source /home/x/y/venv/bin/activate && python asdf.py" >> /tmp/asdf.txt 2>&1"""], shell=True)

pass bash script args as named parameters to a command inside the script

I have a bash script that takes two parameters. Inside that script, I need to call ssh using a heredoc and call a method that expects the two arguments. For example:
ssh -o "IdentitiesOnly=yes" -t -i $key -l user localhost << 'ENDSSH'
/my_python_app.py -u -t tar -p $1 -f $2
ENDSSH
key is set by my script, I know that part is good.
However, my_python_app prints out args and it doesn't show any arguments for -p and -f
I would call my script like
my_script /tmp filename
I use argparse in my python app, but I am also printing out sys.argv and it gives me:
['my_python_app.py', '-u', '-t', 'tar', '-p', '-f']
Note there are no values received for -p and -f. (-u is a flag, and that is set correctly).
How do I pass $1 and $2 to my_python_app as the -p and -f values?

Remove the quotes around the here-document delimiter (i.e. use << ENDSSH instead of << 'ENDSSH'). The quotes tell the shell not to expand variable references (and some other things) in the here-document, so $1 and $2 are passed through to the remote shell... which doesn't have any parameters so it replaces them with nothing.
BTW, removing the single-quotes may not fully work, since if either argument contains whitespace or shell metacharacters, the remote end will parse those in a way you probably don't intend. As long as neither argument can contain a single-quote, you can use this:
ssh -o "IdentitiesOnly=yes" -t -i $key -l user localhost << ENDSSH
/my_python_app.py -u -t tar -p '$1' -f '$2'
ENDSSH
If either might contain single-quotes, it gets a little more complicated.

The more paranoid way to do this would be:
# store these in an array to reduce the incidental complexity below
ssh_args=( -o "IdentitiesOnly=yes" -t -i "$key" -l user )
posixQuote() {
python -c 'import sys, pipes; sys.stdout.write(pipes.quote(sys.argv[1])+"\n")' "$#"
}
ssh "${ssh_args[#]}" localhost "bash -s $(posixQuote "$1") $(posixQuote "$2")" << 'ENDSSH'
/path/to/my_python_app.py -u -t tar -p "$1" -f "$2"
ENDSSH
If you know with certainty that the destination account's shell matches the local one (bash if the local shell is bash, ksh if the local shell is ksh), consider the following instead:
printf -v remoteCmd '%q ' /path/to/my_python_app.py -u -t tar -p "$1" -f "$2"
ssh "${ssh_args[#]}" localhost "$remoteCmd"

ssh restrict commands using Python

I am using a python script to restrict the commands usage using the command argument in the authorized_keys file.
command:
ssh host-name bash --login -c 'exec $0 "$#"' mkdir -p hello
My script is performing required actions to restrict the commands. After filtering, the python script does sys.exit(1) for error and sys.exit(0) for success. After the return value the above ssh command is not getting executed at the end. Is there something else I need to send from the python script to SSH daemon?

The command modifier in the authorized_keys is not (only) used to validate the users command, but that command is run instead of the command provided by the user. This means calling sys.exit(0) from there prevents running the user-provided command.
In that script, after you validate the command, you need to run it too!

I think changing it to
ssh host-name bash --login -c 'exec $0 "$#" && mkdir -p hello'
should do the trick, otherwise bash will assume only the part in the single quotes is the command to execute.
If the second part should be executed even if the first part fails, replace the && with ;

Not getting prompt after executing python script from bash

I am executing a python script from bash script as follows
"python -O myscript.pyo &"
After launching the python script I need to press "enter" manually to get prompt.
Is there a way to avoid this manual intervention.
Thanks in advance!

pipe a blank input to it:
echo "" | python -O myscript.pyo
you might want to create a bash alias to save keyhits: alias run_myscript="echo '' | python -O myscript.pyo"

placing wait after the line to run the process in background seems to work.
Source:
http://www.iitk.ac.in/LDP/LDP/abs/html/abs-guide.html#WAITHANG
Example given:
!/bin/bash
test.sh
ls -l &
echo "Done."
wait
Many thanks

Strange behavior redirecting to a file while using runuser

I'm currently executing a python file with runuser and redirecting the output to a file. This is the command:
runuser -l "user" -c "/path/python-script.py parameter1 > /path/file.log &"
This run correctly the python script but creates an empty log file. If I run without the redirect:
runuser -l "user" -c "/path/python-script.py parameter1 &"
runs correctly the python script and make all output from python script to flow the screen. The output from the python script are done with print which output to stdout.
I don't understand why the output from the python script is not dumped to the file. File permissions are correct. The log files is created, but not filled.
But, if I remove the "parameter1", then the error message reported by the python script is correctly dumped to the log file:
runuser -l "user" -c "/path/python-script.py > /path/file.log &"
The error message is done with print too, so I don't understand why one message are dumped and others not.
Maybe runuser interprets the "parameter1" as a command or something. But, the parameter is correctly passed to the script, as I can see with ps:
/usr/bin/python /path/python-script.py connect
I've tried adding 2>&1 but still don't work.
Any idea ?

Encountered similar problem in startup scripts, where I needed to log output. In the end I came up with following:
USR=myuser
PRG=myprogrog
WKD="/path/to/workdir"
BIN="/path/to/binary"
ARG="--my arguments"
PID="/var/run/myapp/myprog.pid"
su -l $USR -s /bin/bash -c "exec > >( logger -t $PRG ) 2>&1 ; cd $WKD; { $BIN $ARG & }; echo \$! > $PID "
Handy as you can also have PID of the process available. Example writes to syslog, but if it is to write to the file use cat:
LOG="/path/to/file.log"
su -l $USR -s /bin/bash -c "exec > >( cat > $LOG ) 2>&1 ; cd $WKD; { $BIN $ARG & }; echo \$! > $PID "
It starts a new shell and ties all outputs to command inside exec.