[ejabberd] Improvements in ejabberd start script - this proposal needs review

Sander Devrieze s.devrieze at pandora.be
Sat Aug 25 03:45:40 MSD 2007


2007/8/25, Badlop <badlop at gmail.com>:
> > >     * Verifies permissions and the existence of ejabberd.cfg
> > >
> > > Not much interesting: make install will already create dirs, seet
> > > permissions, etc.
> >
> > My idea was to make it more foolproof:
> > * user accidentially deletes an important file or changes it to
> > insecure permissions
> > * system accidentially deletes an important file or changes it to
> > insecure permissions because of a bug
>
> Those reasons are not convincing for me. If there is a trouble-maker
> admin or a buggy system, it's not our task to fix their mistakes.

I'm pragmatic: a trouble-maker admin or buggy system can cause a lot
of difficult to answer questions people send to us. Why not reduce the
numbers of difficult to answer questions before they come to us by
adding some checks in the script?

> Do you know this joke?
> Computing is a race between the Universe and Engineers.
> Engineers trying to develop better idiot proof software,
> and the Universe trying to make better idiots.

nope

btw, this quote is relevant:

" It was during this conference that we realised that the work we were
doing on Erlang was very different from a lot of mainstream work in
telecommunications programming. Our major concern at the time was with
detecting and recovering from errors. I remember Mike, Robert and I
having great fun asking the same question over and over again: "what
happens if it fails?"— the answer we got was almost always a variant
on "our model assumes no failures."We seemed to be the only people in
the world designing a system that could recover from software
failures..." (from:
http://patricklogan.blogspot.com/2007/08/bits-of-wisdon-hopl-iiis-history-of.html
)

> > ==>more security, no strange bug reports sent to the ejabberd project
> > because of stupid errors by this user (the user will get a good
> > description of the problem in the error message so he don't have to
> > search for what he did wrong).
>
> This argument convinced me: if ejabberd can't work, let's detect it
> the sooner the better, and let's report an error message as
> explanatory as possible.
>
> In summary, at last I add your file and permission detections to the
> proposed patch.

ok, so I hadn't had to write down all arguments above, anyway, I leave
them because they are more elaborated versions.

<snip>

-- 
Mvg, Sander Devrieze.


More information about the ejabberd mailing list