exit on a script

[someone_4]

When and why and should you type exit at the end of a script ?

Thanks
Richard

[Rainer von Bongartz]

passing a parameter with exit (exit 2 i.e.) you can check the exit starus from your script. useful if you want to check wheter your script dit an normal exit or aborted on an error condition .

[Ravi_8]

Hi,

after executing the script exit(1) just comes back to prompt without closing the file, where as exit(2) closes the file and comes out

[someone_4]

so i should use exit(2) on every script?

Richard

[Ravi_8]

Hi,
sorry for mistake, it's actually it is exit(0) and exit(1). yu can use exit(1) at the script end.

[James R. Ferguson · Accepted solution]

Hi Richard:

First, it is good practice to exit a script with a defined return value. If you don't, the exit value will be the return value of the last command executed, by default.

Standards define three exit values, the remainder being a user's choice:

0 => command or process completed successfully
1 => command or process failed
2 => command or process had warnings

A good example of a command following these guidelines is 'grep'. An example of a command that deviates is 'diff'. Take a look at the man pages for both to see!

Thus, even these guidelines are not hard-and-fast.

Regards!

...JRF...

[Thierry Poels_1]

hi,

in most cases you don't really need an exit statement at the end of your script. The shell process will terminate at the end of your script unless you have set "ignoreeof" option.
Exit is mostly used to exit from within your script at certain conditions: error, file not found, execute not required, .....
"Exit" or "Exit 0" is normally used to leave the script, and state that is ended normally.
Exit codes >= 1 are normally used to indicate some error has occured.
regards,
Thierry.

[Bill Hassell]

There's one good reason to add exit (or exit 0) at the end of every script: It documents the true end of the file. If someone miscopies or truncates a script, there is no way to know whether all the lines are still there. By having both a comment and exit 0 such as:

# end of script
exit 0

then the reader will be assured that the end of the file is present.

[James R. Ferguson]

Hi Bill:

I can't resist saying "AMEN!". The notation of "end " is so desirable...on reports, in code... At least in languages like ALGOL and PASCAL we have the enforced "END" to delineate the outer block! ;-)

Regards!

...JRF...

[A. Clay Stephenson]

Hi Richard,

There is one more reason, when your script calls another scrip, there is a reliable means of checking the status.
e.g.
XXX=`somecommand.sh`
STAT=$?

Now $STAT is the exit status of somecommand.sh while $XXX contains its output.

This also should be extended to functions within your scripts.

somefunc()
{
FUNCSTAT=0
echo "This is somefunc output \c:"
while [ $# -ge 1 ]
do
echo "Arg ${1} \c"
shift
done
echo
return(${FUNCSTAT})
}

XXX=`somefunc 111 222`
STAT=$?

Here the same rules as above apply. Typically, you test the exit(return) status before ever doing anything with the script(function) output.

The last point is that if you do c programming the same rules apply: main() should also return a value.

Clay

[A. Clay Stephenson]

Hi again Richard,

Looking over the earlier replies, it appears that ravi was not quite correct. There is no default behavior associated with exit(1) or exit(0). Simply follow the standard rule:
0 means good; anything else not so good.