Hello,
Luca pointed out that
> 2) BASH (list) exit status is not taken into account by `set -e`
>
> We use the second every day (not only in the pkg-fso stuff), thus
> clearly understanding how it works (especially WRT the docs) is a win
> for everyone.
damn right. I just played with it a bit. It seems like a bash-bug or at least a
bash-inconsistency with regard to what set -e interprets. The (list) have a
return value, so the man page says, that is the exit status (Luca quote that
before).
In consequence, the ()s should be treated like a function that has a return
value.
The script
set -e
retOne {
return 1
}
retOne
echo "will not be reached"
will not reach the echo line. Also
if retOne; then
echo "will not be reached"
else
echo "will be reached"
fi
will always chose the else branch, and is behaving just like
if (cat /dontexist); then
echo "will not be reached"
else
echo "will be reached"
fi
So, to have the echo reached in
set -e
(cat /dontexist)
echo "should not be reached but is"
seems inconsistent.
Another issue is that
set -e
(cat /dontexist; echo something)
does execute the echo. And so does
(set -e; cat /dontexist; echo something)
This should not happen, the command execution environment should have inherited
the
settings from its parent. At least
(cat /dontexist && echo something)
works as expected :)
The page
http://tiswww.case.edu/php/chet/bash/bashtop.html#Bugs
explains to use the tool bashbug to report bugs. If it
is a bug. Luca?
Many greetings
Steffen
_______________________________________________
pkg-fso-maint mailing list
[email protected]
http://lists.alioth.debian.org/mailman/listinfo/pkg-fso-maint
