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

request for clarification wrt POSIX /bin/sh and local



hi -devel and -policy,


this hasn't been discussed in a while, and i'm wondering if we could
come to an agreement to clarify policy on the matter.  currently, policy
states:

currently, policy (10.4) states:

	...The standard shell interpreter /bin/sh can be a symbolic 
	link to any POSIX compatible shell, if echo -n does not 
	generate a newline. Thus, shell scripts specifying /bin/sh as 
	interpreter must only use POSIX features...

however, many important parts of debian (lsb and debconf are the two
that i've heard mentioned) also rely on the existance of the non-POSIX
"local" construct to declare local variables in the scope of a function.

most modern shells (bash, dash, ash at least) provide support for this,
but the strictly POSIX posh does not.  

over two years ago, a bug[1] was opened against debian-policy asking for
an extension to the POSIX related exceptions to include local support
along with "echo -n", which went largely unanswered.    however, the
topic has been discussed a few times since then[2], and as one thread
evolved, it seems like some of the policy maintainers agree with the
rationale for modifying policy[3]

in the meantime the release managers decided that packages depending on
such constructs were at worst non-RC policy violations.  because of this
decision, and also the widespread use of the feature, it seems to me
that policy should be modified to reflect this... however maybe someone
has an argument[4] for why it should not.

but either way, could we get an authoritative answer on this?


thanks,
	sean


[1] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=294962
[2] http://lists.debian.org/debian-devel/2006/11/msg00168.html
[3] http://lists.debian.org/debian-devel/2006/11/msg00317.html
[4] where the argument goes beyond, "well, it's not POSIX" and gives
    the benefits for staying closer to POSIX, that is.

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


Reply to: