Re: Proposed new POSIX sh policy
Thomas Bushnell BSG <firstname.lastname@example.org> writes:
> On Tue, 2006-11-14 at 18:38 -0800, Russ Allbery wrote:
>> I think there are obvious reasons to exempt shell builtins from this
>> requirement, so you're going to have to present more of an argument
>> than this. I think this is a very strained reading of that section of
>> Policy, and in my opinion is "obviously" not what that section is
>> trying to say.
> We have not yet exempted builtins. I am not, in fact, opposed to doing
> so. But I want a list of exemptions to be specific.
I think this would be a great deal of work for little useful benefit.
> Are you seriously saying that *any* shell builtin is exempt? That a
> shell may override *any* Debian behavior?
I'm saying that I don't believe there is anything in Policy right now that
prohibits such behavior. Not every possible bug is proactively ruled out
by Policy. Many are left to the common sense of the maintainers. The
purpose of Policy is to document the correct course of action in which
there are multiple reasonable choices or where we need a ruleset in order
to produce an integrated and consistent distribution.
I personally don't see a current need to state in Policy "don't add shell
builtins that override common commands without providing the features of
those commands"; I think that problem can be adequately dealt with by
normal bugs and the application of common sense.
>> So every shell in Debian is buggy, since none of them provide all the
>> functionality in all of their builtins that coreutils provides.
> Yes, I'm happy to say that. These are not crucial bugs; Debian has lots
> of bugs. :) Not every bug is of critical importance, but yes, these are
Okay, in that case I believe we have a fundamental difference that I don't
think spending further time discussing is likely to resolve.
I disagree with your proposal and do not believe it is the direction we
> Your proposal seems to me to be even worse. Mine says that there are
> lots of low-grade unimportant bugs hanging around, which probably don't
> need lots of attention except when cases arise where they cause real
> problems. Yours says that there is no bug at all in a shell which
> decides to override the behavior of debconf in random and unpredictable
The absence of a requirement in Policy is not a positive declaration that
something is not a bug. I do not want to document every possible bug in
Policy; I don't believe that's a useful use of resources. No one has to
my knowledge uploaded a shell intended for use as /bin/sh that overrides
debconf in random and unpredictable ways, and I believe the likelihood of
anyone doing that and then being sufficiently convinced of its technical
merits as to demand citing chapter and verse of Policy to resolve the
situation is exceedingly low.
Russ Allbery (email@example.com) <http://www.eyrie.org/~eagle/>