Re: Proposed new POSIX sh policy
Thomas Bushnell BSG <firstname.lastname@example.org> writes:
> Russ's patch is no good, at least, it does not address the problems I
> have raised in the past.
> A Posix shell is allowed to have a builtin for ANY command without
> restriction, and as long as the builtin has the behavior specified by
> Posix for that command, it is a "Posix compatible shell."
That's correct; I'm not attempting to address this problem in this
proposal. This is a problem with the current Policy wording and continues
to be a problem with the new Policy wording, so this change is not a
regression in this regard. I'd prefer to separate the question of what
shell features need to be supported from the question of how to handle the
general built-in issue, since I think there's more project consensus
around the former than around the latter. The only overlap in this case
is with the test command; local is not a meaningful external command, and
the test behavior required by my patch is a subset of what would be
required under your proposal.
> I have proposed as an alternative that we say that a shell is suitable
> for Debian's /bin/sh if its builtins behave just as do the Debian
> programs installed in the standard paths, and then we can make a
> specific set of exceptions for builtins that we don't care if they
> behave the same way as Debian's versions.
My impression of the previous Policy discussion was that there was not a
consensus around this change, so I'm trying to reach a consensus around a
simpler incremental change that deals with one problem (while still
leaving others opened). This should in no way be taken as a cutting off
of debate of the larger issue.
Russ Allbery (email@example.com) <http://www.eyrie.org/~eagle/>