Setting environment variables in a shell script

[Glitz]

I tried to temporarily set environment variables in a bash shell script. However, when the script terminates so does the shell and the environment variables are lost. Is there anyway to transfer them to the original running shell?

 

Glitz.

[liquidzoo]

put them in the bashrc file maybe?

[fuzzylizard]

Theortically, I think there is a way of doing that, however, if you want to keep them long term then you need to put them into the .bashrc or .profile file. This should ensure that they are loaded each time you log in.

[Qchem]

don't start the script with #!/bin/bash and don't finish it with exit 0, then just run the script as

 

. /path/to/script

 

it works for me [tm]

[ramfree17]

Qchem's method is more commonly known as sourcing a script file. it does not invoke a separate bash instance to run the script thereby all changes are retained by the current (bash) interpreter. you could do the same by calling

$ source [scriptfile]

 

always remember that shell variables are not normally propagated to child processes unless they are explicitly exported (EXPORT <variable>). the other way around is impossible as a child process cannot transfer their environment variables back to their parent process.

 

i think this information is on the bash how-to in tldp.org. :)

 

ciao!

[Glitz]

Thanks guys!

 

I put the exported variable in the /etc/profile file. Works great!

 

Now I have the latest Glade-2 and Anjuta stuff compiled and operating on LM9.1 at work.

 

Glitz.