Re: set -e is ineffective for subshells

[Paul Jarc]

Chet Ramey <chet@nike.ins.cwru.edu> wrote:
>> Repeat-By:
>>         bash -e '(exit 1); echo still here'
>
> POSIX.2 says that `set -e' causes the shell to exit if a simple
> command fails (subject to a few irrelevant conditions).  A subshell
> like `(exit 1)' is not a simple command.

This behavior is incomatible with traditional sh, and it makes it
harder to write correct, portable #!/bin/sh scripts.  Does POSIX
prohibit exiting in this case, or merely not require it?


paul