Re: #! -- reconsideration?
- To: email@example.com
- Cc: Cragun@cise.ufl.edu, Don.Cragun@Eng.Sun.COM, Don@cise.ufl.edu, firstname.lastname@example.org, email@example.com, firstname.lastname@example.org, email@example.com, firstname.lastname@example.org, email@example.com, firstname.lastname@example.org, email@example.com, firstname.lastname@example.org
- Subject: Re: #! -- reconsideration?
- From: Ted Baker <email@example.com>
- Date: Thu, 11 May 2000 08:14:44 -0400 (EDT)
- Message-id: <200005111214.IAA20667@dad.cs.fsu.edu>
| > Or, pick something else, like your favorit cartoon expletive
| > #*?&!#@*#!.
| If (and only if) a new syntax/mechanism have to be defined, could
| be a good idea to explicitly separate in the syntax itself
| the standard body from the version from the command/interpreted executed ?
That is what I had in mind, i.e., to separate the issue of how
to get the standard POSIX shell from the more general issue of
how to support various other interpreters (Perl, etc.). The latter
interpreters are pretty well outside the scope of POSIX.
Using a special syntax that cannot be interpreted as a path name
to denote a POSIX-compliant shell would eliminate the possibilty
of conflict with a given existing implementation or application's
assumptions about the meaning of any particular pathname and would
eliminate the issue of whether the system even supports conventional
This would require a small addition to the loader, to recognize
the special syntax, and would only be useful to new applications.