Beyond shar (Re: shars and security concerns.)

[Peter da Silva]

In article <15462 at bfmny0.UU.NET> tneff at bfmny0.UU.NET (Tom Neff) writes:
> With proper standardization it should not be necessary for an unshar
> program to handle all kinds of weird shars.

I think that optimistic. Consider all the companies maintaining backwards
compatible bugs in their software because people ignored *enforcable*
standards!

[ I said... why not Software Tools format? ]
> Because it doesn't self extract.

Neither does a shar on anything but UNIX, nor a DCL-shar on anything but
VMS, etc...

> Also, my memory may be faulty but the last time I looked at ST it seemed
> to me there was no provision for recording modes or ownership.

Modes? You mean like Read, Write, Extend, Delete, Execute, Set-User-ID,
Set-Group-ID, etc...? Or do you mean like Stream-LF vs Fortran Cariage
Control vs Fixed Record versus Card Image etc...?

Ownership? What's that? Or does that relate to ACLs...?

Things like that are inherently O/S specific, and if they're important
include a shell script or DCL command file to set them. That can also be
automatically generated.

How about making everything not between a -h- and -t- a comment, and make
the file look like this:

sed -e 1d -e s/^X// > file << '-t- file'
-h- file Sun May  6 17:09:59 CDT 1990 ./file
Xstuff
Xand more stuff
-t- file
etcetera...

As for coded comments...

I much prefer having an automatically-generated shar-mode.sh file appended
to the archive, like the MANIFEST is in some sharchives right now.
-- 
`-_-' Peter da Silva. +1 713 274 5180.      <peter at ficc.uu.net>
 'U`  Have you hugged your wolf today?  <peter at sugar.hackercorp.com>
@FIN  Commercial solicitation *is* accepted by email to this address.