[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Proposed new POSIX sh policy

> > I know what Posix.2 says, but it does not define the term "POSIX
> > compatible shell".  Can you tell me what that means?  I really am
> > genuinely stymied.  I think some people have an incorrect
> > understanding of what POSIX actually says in this regard, but I'm
> > not sure.
>         If you truly are stymied by the phrase "POSIX compatible
>  shell"  in reference to ISO/IEC n9945-2:1993 and addenda, where it
>  defines the command interpreter, or shell, then truly your mind set
>  is not attuned to the requirements of this mailing list.

Instead of insulting me, can you just tell me what it means?  I have
read and understood the standard since it was first issued.

>         If you truly do not understand the meaning of the requirement
>  that your maintainer shell scripts  must e restricted to the dictates
>  of standard shell features as defined by POSIX, with the additiosn of
>  echo -n and test -a/-o and a few things like that, feel free to ask
>  on -mentors for help. I am sure there will be people who have no
>  problems interpreting policy for you there.

I don't think you understand; and I think you are assuming that I'm just
trying to be annoying.  I'm sorry you think this, and it's insulting
that you would think this, but I'll try to ignore the insult.

POSIX does not distinguish at all between the status of echo, ls, and
test.  It puts them in the same section, talks about them in the same
terms, and so forth.  In no way does POSIX say or imply that the
behavior of "test" is a "shell feature" and the behavior of "ls" is not.

So can a Debian shell script use fileutils-specific parts of "ls"?  If
the answer is "sure", then why is "test" not the same?  Where in Policy
is this explained?

POSIX allows the command interpreter to extend the specification, of
course.  It does not define whether a given extension is a shell
extension or not.  The addition of a command, and the addition of a new
option for an existing command, are treated in exactly the same terms,
for all of the commands in the bulk of the document (where ls, echo, and
test are specified, for example).  Nothing about Posix says that an
extension to echo is a "shell" thing, but an extension to ls is not.


Attachment: signature.asc
Description: This is a digitally signed message part

Reply to: