Martin Vaeth <[email protected]> wrote:
> Joerg Schilling <[email protected]> wrote:
> >
> > bash vs. POSIX, as bash tried to ignore long existing
> > rules just because the bash maintainer did not understand them.
>
> Are there really several? I know only one such example:
One is that "sh -ce cmd" did not exit on error for some kind of commands.
This is where I have been able to convince the bash maintainer together with
David Korn for bash-4.0. This was a nightmare for make.
Another is e.g. that bash makes job control for commands in scripts or commands
from "sh -ce cmd". This is another nightmare for make, as this prevents layered
makefiles from terminating when ^C is typed as some comands run in different
process groups. Smake for this reason has a special autoconf test for /bin/sh
being bash and tries to manually kill(2) the dependencies if they are run by
bash.
> bash insists on compound commands ("{ ... }" or "( ... )")
> for the function body while according to POSIX also
> non-compound commmands can form the body, e.g.
>
> hello() echo "hello world"
>
> is a valid function definition according to POSIX
> (and thus works in dash or also zsh) but not in bash:
> Rumors say that the bash maintainer intentionally
> excluded this due to some misinterpretation of the
> POSIX formulation.
This is valid in the Bourne Shell already, so it is something that can be seen
as very basic.
Jörg
--
EMail:[email protected] (home) Jörg Schilling D-13353 Berlin
[email protected] (work) Blog:
http://schily.blogspot.com/
URL: http://cdrecord.org/private/
http://sourceforge.net/projects/schilytools/files/'
