I'm still confused what you are trying to actual do here.

I still think you would be better sourcing your file from you script.

Let me put it like this:-
my_env.ksh

Code:

#!/bin/ksh
a=Hello
b=Robin
c="$a $b"
echo "$c"

my_script.ksh

Code:

#!/bin/ksh
echo "Starting my script."
echo "$c"
. my_env.ksh          # I'm assuming that this is in your $PATH
echo "I have set up the environment within this shell."
echo "My name is $b"
echo "$a world!"

Then I can run

Code:

$ echo $a

$ my_script.ksh
Starting my script.

Hello Robin
I have set up the environment within this shell.
My name is Robin
Hello world!
$

You can see that the values of $a, $b & $c are available to your calling script. They way you have it with eval leaves you open to odd things, such as your echo with quotes. Calling

Code:

eval echo "echo "this is line 1"
                 date"

.... will just confuse things because it will try to match up quote etc. , so you are actually going to execute:-

Code:

echo echo this is line 1"
"

which will work, but is a bit odd. How would you cope with multiple / escaped quotes? Consider odd_file

Code:

echo "This is the 
\"text\"
...I think"

An eval of that might go a bit wild. If you run this on the command line:-

Code:

$ while read line
> do
>    eval $line
> done < odd_file
This is the
ksh[3]: text:  not found
ksh[3]: ...I:  not found

Where you could:-

Code:

$ ksh odd_file
This is the 
"text"
...I think

or

Code:

$ . odd_file 
This is the 
"text"
...I think

Perhaps I am being stupid and not seeing what you have explained. Can you be very direct as to the purpose. I don't mind how terse you make the language, I'm just very confused.


Robin