Re: bash exorcism experiment ('bug' 762923 & 763012)
On Sat, 11 Oct 2014, Theodore Ts'o wrote:
> I assume that posh meets the strict definition of 10.4. And so
> without actually changing policy, someone _could_ try setting /bin/sh
> to be /bin/posh, and then start filing RC bugs against packages that
> have scripts that break. Yes?
Yes, modulo two things:
① Bugs in posh – posh derives from pdksh, currently.
It probably would need to be rebased on mksh. I wonder if
it’s worth doing it upstream; currently, there is not enough
interest for it from actual users. (I did hear from one or
the other distro they’d prefer a /bin/sh that did not have
extensions over POSIX, but…)
② The CTTE decision of #539158 to not overturn Md, which in
turn requires for a shell to have a printf(1) builtin since
#428189 (the only way to fix it that does not involve Md)
is also, de facto, rejected – meaning a switch to posh will
cause most systems to fail to boot.
(The mksh binary package ships an lksh binary, which uses
the C “long” type for arithmetics, as required by POSIX,
instead of consistent arithmetic ops, and also bundles a
minimal (busybox/klibc-like) printf builtin, in order to
be able to use it as /bin/sh on Debian.)
15:41⎜<Lo-lan-do:#fusionforge> Somebody write a testsuite for helloworld :-)