Shell Programming and Scripting

Thanks for the info re ksh ... always nice to learn new things :-)

My point was there is a difference between calling an executable shell script directly (./myscript.sh) and relying on the hash bang line to determine what shell to run it on (the current shell of course if no hash bang line is present) and calling it using a specific shell (ksh myscript.sh).

Under ksh vino's example works fine (using ./myscript.sh). Unless of course you call it using 'ksh myscript.sh' (and circumventing the hash bang line). Which was why I was trying to determine how it was called.

I think this is the mechanism to provide command level atomicity.
If you change the script in between it just work as per
new set of commands .

Just out of curiosity.

What changes ? What will the final script look like ?

I mean if you don't delete the the file . For example
if the shell script is like this

#!/bin/ksh

echo "Hello"
sleep 100
echo "World"

You run the command the script like say test.ksh

Then from other window change the script to look like this
during the program execution .

#!/bin/ksh

echo "Hello"
sleep 100
echo "World"
echo "Testing "

The command executed in first session will display will echo the text testing as well

What I meant here is this ... Statement level atomicity like in database is also maintained in the command level at os .
I probably read something about this during my school days ... may be the book was authored by stephen prata

Thanks
Ashok

There is no "statement level atomicity". You cannot modify a running script and have any assurance that anything you expect might happen. ksh reads in buffers at a time until it has enough to run. When it needs more, it will read another buffer. It often has unprocessed input sitting in a buffer. It almost always splits a statement as it reads a buffer. It has a file pointer into the script. Like any other file pointer, it is a character position, not a line number.

That was a highly informational series of posts-thanks every one. I didnt execute the script as "ksh remove.sh" but I had executed it as ". remove.sh" (which didnt work out). But executing it as "./remove.sh" works fine.

Code:

. remove.sh

and

Code:

ksh remove.sh

are very similiar in the sense that they are passed as a parameter.

To

. - the current shell
ksh - a new ksh subshell.

vino