[mdlug] I need a script Hero

[Raymond McLaughlin]

Aaron Kulkis wrote:
> Raymond McLaughlin wrote:

>> I agree that file extensions, while optional, are often a good option.
> 
> But not on scripts that are meant to be commands....

Is there some practical reason for this, some sort of convention, or is
it just your own subjective view?

My own, admittedly subjective, view is that I like to know at a glance
the file type of a command: I view it as a courtesy and a convenience.
I don't see what is gained by obscuring it to save three bytes in the
file name.

> Filename extensions, on the other hand, should really only be used
> for files associated with compilers (source files, non-executable
> object files, and libraries).

And I suppose orange juice is just for breakfast. :)

Actually it's not unusual to see something like:
  A command on the $PATH symlinked to a script
    /usr/bin/firefox --->/opt/mozilla.org/firefox/firefox.sh
  the script checks and sets up the local environment then calls the
  binary. ie. firefox.bin (or some such).
This "division of labor" is similar to that used in your example, where
you might have files: 'program.c', 'program.o' and finally 'program' in
the same directory. I realize I'm using here an extensionless file name,
but in this example it is a binary.

I feel pretty sure that this is a case where there is no one right
answer, especially not one that fits all situations. In the case that
originated this thread we were talking about a setup script found in the
root of a CD. Whether you read the directions or just mount the CD and
pull up a dir, you really can't miss it, and I think the extension is
appropriate. I cases where the command is buried somewhere in the system
tree then the presence or absence of an extension would depend on how
you expect the user to even know about the command. And if there is any
question a symlink can be used:
    setup ---> setup.sh

Can you name an instance where the presence of an extension caused a
problem? Particularly a problem that wasn't really a problem of poor
documentation?

Would like to know
Raymond McLaughlin