Re: Uploaded faqomatic 2.506-2 (source all) to master
Hi,
>>"Richard" == Richard Roberto <robertr@nwmarkets.co.jp> writes:
Richard> On Tue, 10 Feb 1998, Christian Schwarz wrote:
Richard> fom is already a small stub. Making it a much larger, more
Richard> complex script just to meet an esoteric "conf file" policy
Richard> seems rediculous.
quite. Adding two lines, however, is not.
Richard> Besides it needing to untaint any external
Richard> data, it may also need to run a different binary, depending
Richard> on if you use a standard perl binary or the debian method.
Richard> This is for people who want (or need) to run the script setid
Richard> (or run a setid embedded interpreter). Am I off base on
Richard> this?
I think so. In that case all code needs to ``untaint''
configuration data it gets from /etc/*, which is not world
writable. The law of diminishing returns (after all, it is not as if
this were VMS ;-)
If it needs a different binary, or a different language, be
design, there should be another cgi bin script. If the config makes
the program fail, the config is in error. (I can set up my
sendmail.conf to send all mail to president@whitehouse.gov. Whose
fault is that?. I can add a line that says exit in /etc/profile too.)
Richard> Perhaps the script isn't what needs tweaking, but rather the
Richard> policy?
I do not agree. This is making a mountain out of a mole
hill. I like having all configurable things under /etc (makes
backups easier [I do a incremental backupo weekly, but every other
day for /etc and /home/srivasta])
Richard> There are numerous scripts on my system in the /usr
Richard> hierarchy. As system administrator, I reserve the right to
Richard> modify any one of them. Does that also break debian policy?
Kinda. I just move them to /usr/local/scripts, which is
earlier in my PATH, and tweak them there. But hey, it is your systm,
break any rules you want ;-)
As a developer, you can't go about breaking policy.
Richard> What I mean is, should all scripts be listed as conf files?
No. Scripts are not meant to be tweaked in place. If you
wanna, copy them elsewhere. Or break policy locally.
Richard> I think this is an old discussion. It would be better if we
Richard> had a separate "script" tag. If there are security
Richard> implications for any part of this discussion, then I think we
Richard> need Ian's input.
So the rest of us are just munchkins? ;-) A policy does not do
much good for security on your machine. It just makes us more
consistent, and one may feel warm fuzzy knowing that all local config
is confined to /etc.
Richard> In any case, it would be better to leave fom the way it is,
Richard> unless the upstream author agrees to change it. He's a
Richard> pretty agreeable guy, so maybe someone should ask for his
Richard> input on it.
I disagree. We change pacxkages all the time to conform to
debian policy. Thats what systems integration entails. That is what
sets us apart. We have a policy, and we follow it.
manoj
--
A deed is not well done if one suffers after doing it, if one bears
the consequences sobbing and with tears streaming down one's face. 67
Manoj Srivastava <srivasta@acm.org> <http://www.datasync.com/%7Esrivasta/>
Key C7261095 fingerprint = CB D9 F4 12 68 07 E4 05 CC 2D 27 12 1D F5 E8 6E
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org .
Trouble? e-mail to templin@bucknell.edu .
Reply to: