I'm working on a script (to be executed via bash) which must set a variety of file permissions and ownership. Because this script may change file ownership, it must have elevated permissions.

I'd like for the script to modify the permissions of files using pathnames relative to its location on the file system. For example if permissions.sh is in scripts and target-files contains the files to be modified:
I don't want to have to refer to the full path names of all these files because they may be in /var/www on one machine and /var/www/project-x on some dev's workstation.

I was thinking I could use relative path names (e.g., ../target-files/targe1.php) but I expect this would be evaluated relative to the cwd of the executor rather than the location of permissions.sh.

I unfortunately have little bash experience setting variables and string operations so I don't know how to capture the directory of permissions.sh from within that script and then concatenate filenames, etc.

Any help would be much appreciated.

In the script, change the directory to the location of the script file. Then write your commmands relative to that point.

1 members found this post helpful.

Thanks for your response.

Having looked at your link and searched a bit, I've concocted a brief bash script, test.sh, that will cd to the parent directory of the directory where the script is located:
I expect that I could then simply enter my chown/chmod commands with relative pathnames and it would probably work. However, I'm wondering what effect that cd command might have. Suppose I run this test.sh script and then run subsequent scripts or commands from some other script, other.sh ?
I've tested this on my workstation (Ubuntu 12.x) and it appears that the cd commands only last for the duration of the current script and do not affect the cwd in other.sh or in script1.sh and script2.sh. Can I reasonably expect this to be the case on all *nix platforms? E.g., CentOS?

directory changes... stay with the current shell and are inherited by child or sub shells.

Let's say I make a script with the following to demonstrate:

Assuming when I run the script that my $PWD is /,
I would expect this output:

Which ends up being correct.
Examine that and why each one is as it is. Specifically, why the last one != the second to last one (remember, going into a child shell will cause the child shell to inherit the current directory but discard any changes once it exits)
A really easy way to test this is by typing this in the terminal:
Your current directory doesn't change, yet you know it did change directory in the subshell since it listed /'s contents.

You are most likely using bash, so this is bash behavior. If other *nix platforms use something besides bash, this behavior could change.
However you could easily load up a bash shell (type in bash in the shell)
OR - type #!/bin/bash
This indicates to the script reader that it should be interpreting it as bash script (my understanding of this is less clear.)

Since starting a script opens a child shell, this is following that behavior.

1 members found this post helpful.