Script interpreter locations

Jakob 'sparky' Kaivo jake at nodomainname.net
Mon Mar 22 11:38:12 PST 1999


On Mon, 22 Mar 1999, Daniel Quinlan wrote:

> Jakob 'sparky' Kaivo <jake at nodomainname.net> writes:
> 
> > Should there be a standard place to put script interpreters (e.g. perl,
> > tcl, python)? There is a thread on the kernel mailing list about a patch
> > to support putting things like:
> >
> > #!perl
> >
> > instead of the full path to perl in scripts, but it really shouldn't be
> > handled in the kernel. Perhaps something should be specified like:
> 
> Exactly what problem is someone trying to solve?  There is no problem.

The guy who posted it was trying to make broken scripts run.

> Handling this in the kernel is stupid.  Worse, it will only work if
> your script is running you have one platform.

Which is what I told him.

> There is a standard place (for perl, sh, csh, etc.) specified by both
> FSSTND 1.2 and FHS 2.0.  perl is /usr/bin/perl.

I wasn't aware of this. I guess I should RTFM a little more closely next
time. ;)

> > If the script interpreter foo is provided, the location of the interpreter
> > will be /usr/bin/foo.
> 
> That's about how it's phrased.
> 
> > Or something similar, so that scripts that depend on an interpreter not
> > part of the base, there is still a standard place for it.
> 
> We don't want to extend the base, so FHS is the place to keep the
> definition.
> 
> - Dan
> 

+-----------------------+-----------------------------+--------------------+
| Jakob 'sparky' Kaivo  |    jake at nodomainname.net    |       jake at ndn.net |
| NoDomainName Networks | http://www.nodomainname.net | http://www.ndn.net |
| AtDot E-mail Services |    http://www.atdot.org     +--------------------+
+-----------------------+-----------------------------+     whois jkk12    |
| Thanks to advances in shortness, I have updated my  +--------------------+
| ~/.signature. Note that nodomainname can now be replaced with just ndn.  |
+--------------------------------------------------------------------------+



More information about the lsb-discuss mailing list