Re: securety related question...

On Tue, 8 Oct 2002, Erika Pacholleck wrote:
> [08.10.2002] Bill maltby - LFS Related <-- :
> > On Tue, 8 Oct 2002, Erika Pacholleck wrote:
> > > 
> > > I *do* normal user commands, typing at command line,
> > 
> > I missed that. My apologies to you.
> 
> Ok, *smile*. You stated, that the OS is responsible for that mess
> which comes up with commands, when in our example a directory has
> been deleted; and you told me to try some of the utils in such a
> case. I did that. And I found that on the machine I am currently
> working on, the external programs could handle this in way which
> (I express it carefully now) is correctly interpretable if one
> uses the brain.

?? Required for a TDU (Typical Dumb User)?   :-)

> Furthermore bash and its builtins like pwd were producing at the same
> time results which can make you laughing if you know what's going on.
> 
> So from what I saw on my machine I am more convinced than before
> that in our case it is a bash matter and has nothing to do with
> the capabilities of the OS.

Well, I guess we come to an impasse here. I operate on the assumption that
OS-provided services should be expected to operate *correctly* (whatever
that is) and that applications (utilities) should not have to program
around failures of the OS. Since I view the deletion of an *active*
directory as a failure of the OS, I can not blame bash if it does not
handle the situation in a "rational" fashion. However, I can applaud the
diligence of the designers/programmers of those applications/utilities
that do something rational with the situation, although I believe they
should not have to be concerned with those issues.

> All those programs were compiled at the same machine with the same
> linux files at hand. And if one program can deal with a deleted
> directory fairly good and the other one cannot, I blame the programs
> and not the OS.

Impasse achieved. :)

> Note, I am not saying I find such deleting actions of "active"
> directories ok, and maybe one can/must blame the OS for that.
> I only say that the OS is not to blame if other programs can
> not deal with the mess which comes up then, because tests here
> showed that a program *can* deal with that.

That is like blaming the child for the milk spilled by the parent because 
the child did not properly plan to prevent the parent from spilling the
milk. Some children might have prevented it, but that does not mean we
should expect all children to be able to do so.

But, OK. Different views here of what components bear what responsibility.
If the OS ever does what I think it should do, then all the programs will
be operating correctly.

If bash, et al, designed/programmed with this "flaw" in mind, they would
operate "rationally" when this OS "error" occurred, as do some other
utilities.

-- 
Bill Maltby
[email protected]

-- 
Unsubscribe: send email to [email protected]
and put 'unsubscribe lfs-security' in the subject header of the message