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

Re: ash/echo/POSIX/SUS



Hi,
>>"Herbert" == Herbert Xu <herbert@gondor.apana.org.au> writes:

 Herbert> Since the behaviour of echo when -n is the operand is
 Herbert> implementation defined, POSIX compliant scripts must not use
 Herbert> it.

        What does this have to do with ash? I thought ash was a shell
 I could use instead of bash (you know, the readline dependency, etc)
 to replace /bin/sh. If ash gratuitously breaks my system as some
 kinda policy cop, then it is useless as a /bin/sh replacement, and I
 would like a strong reminder placed all over the documentation and
 web sites that ash is not a real shell meant for use on user systems
 as /bin/sh, since it shall from time to time delibrately break your
 system to flush ut bugs.

        If you want to deprecate the use of as, just go right ahead.

 Herbert> * For the release after potato, we need to make a decision
 Herbert> on whether to change the policy to not requrie POSIX
 Herbert> compliance on #!/bin/sh scripts or actually enforcing it.

        Enforcing it is fine; breaking the system is not a good means
 of enforcing policy. 

        Find the scripts, and file bug reports. Release an
 experimental ash; that people _who want to_ may install to flush bugs
 out.

        Delibrately breaking the system by being stricter than you
 have to removes ash as a candidate for /bin/sh on all but toy
 machines. 

 Herbert> We are talking about scripts and our policy.  Currently
 Herbert> those scripts violate the policy.  So one of them must be
 Herbert> changed.

        Bullshit. Find the scripts, and file bug reports. Breaking
 user systems out there is a lousy way of quality control.

 Herbert> I don't think so.  Having an inconsistency of this magnitude
 Herbert> between the policy and reality is a really big worry to me.

        And breaking unsuspecting systems is your solution? Jeez.

        manoj 
-- 
 Don't tell any big lies today.  Small ones can be just as effective.
Manoj Srivastava   <srivasta@debian.org>  <http://www.debian.org/%7Esrivasta/>
Key C7261095 fingerprint = CB D9 F4 12 68 07 E4 05  CC 2D 27 12 1D F5 E8 6E


Reply to: