When setting up a variable in
.bashrc, should I use this?
Or would this be enough?
What is exactly the difference (if there is one)?
only sets the variable for the duration of the script (
.bashrc in this case). Child processes (if any) of the script won't have VAR defined, and once the script exits
VAR is gone.
VAR to the list of variables that are passed to child processes. Want to try it? Open a shell, do
PS1="foo > " bash --norc
The new shell gets the default prompt. If instead you do something like
export PS1="foo > " bash --norc
the new shell gets the prompt you just set.
Update: as Ian Kelling notes below variables set in
.bashrc persist in the shell that sourced
.bashrc. More generally whenever the shell sources a script (using the
source scriptname command) variables set in the script persist for the life of the shell.
Both seem to work just fine, but using export will ensure the variable is available to subshells and other programs. To test this out try this.
Add these two lines to your .bashrc file
TESTVAR="no export" export MYTESTVAR="with export"
Then open a new shell.
echo $TESTVAR and
echo $MYTESTVAR will show the contents of each variable. Now inside that same shell remove those two lines from your .bashrc file and run
bash to start a subshell.
echo $TESTVAR will have an empty output, but running
echo $MYTESTVAR will display "with export"