Re: Implicit linking of local .so

[Mads Bondo Dydensborg <madsdyd@challenge.dk>]

On Thu, 2 Dec 1999, Steve Baker wrote:

> Oh - I see.  'export' is a Bourne-shell thing - so it's bound to
> be in the Posix spec.

My mistake.

> 
> That's what export does.
> 
> Without export, the variable is confined to the current shell
> and is NOT passed on to daughter shells or programs.
> 
> With export, the variable is passed down to daughter shells
> and programs - but not 'up' to parent shells or programs.

Hm. I think we agree here (and you obviously know what you are saying).

> 
> AFAIK, there is no way to make the change in a variable have
> an effect on the script/program that created this shell. That
> would be "A Bad Thing" indeed - which is why UNIX makes it
> (essentially) impossible.

Yes.

What I was trying to suggest - and this may be wrong - is that the
construct

$> VARIABLE=VALUE executable_program

makes VARIABLE available to executable_program, but not the current shell.

I may be seriously misguided. My apologies.

Mads

-- 
Mads Bondo Dydensborg.                               madsdyd@challenge.dk
Faced with the prospect of rereading this book, I would rather have my brains
ripped out by a plastic fork
                 - Charles Cooper, ZD net, in review of B@TSOT by Bill Gates.