Re: make anacron a base package
On Wed, Mar 31, 1999 at 05:05:59PM -0600, John Goerzen wrote:
> > Yes, but person I was replying to was mentioning the exim.conf file.
> > This was just to prove that exim.conf is actually a lot smaller and
> > more appropriate for in place editing.
>
> No, you're not making a valid comparison. You're comparing a file
> used by the end-user (exim.conf) to what is basically an internal data
> file these days (sendmail.cf). You ought to compare exim.conf to
> sendmail.mc.
But all of exim configuration is done in exim.conf (although that is also
definable). It, is equivalent to all sendmail files (while, of course,
aliases for both are being kept elsewhere, in /etc/aliases).
> > > And again, what does this show? Sendmail has been around a LOT longer
> > > than exim. a LOT.
> >
> > I know that very well. Still, new bugs and security flaws in sendmail
> > are being discovered these days. One would think that with software
> > *that* tested, it would have to be more bug-free? Apparently, not with
> > sendmail.
>
> Urmm, wait, how does the fact that people find bugs make it more
> buggy? This is very flawed logic. Perhaps the fact that so very
> many, many more people are using sendmail and poking at it on a daily
> basis means that there are more people to find bugs when they arise?
But do the bugs arise, or were they there in previous releases? That
is the problem. If those important bugs are created by new features,
they would be stopped by our release process. But some of them are
bugs in functions implemented from the very beginning... :(
> > Exim is getting more and more users (and Debian is helping that),
> > but still, it only had one high urgency and one medium urgency upload.
>
> again, this says nothing. Exim has been around a lot less than
> sendmail. Far fewer people are looking for exim bugs.
Still, fair number of people are looking for exim bugs. So far, they
have found a *lot* less important bugs than in sendmail. Apparently,
exim authors have learned from sendmail's experience.
> You haven't even mentioned why there were high or medium-urgency
> uploads for either MTA.
I trust our maintainers, when they declare that they uploaded high or
medium urgency package, that it is not about fixing a typo in a manpage!
I admit, some of these are because of Debian specific issues, but see
these (I've cut out irrelevant parts):
sendmail (8.9.3-1) frozen unstable; urgency=high, closes=33152 32872
-- debian/site.config.m4 -- new defines for DOS fix
-- debian/debian.m4 -- new defines for DOS fix, bump version
* New upstream version - bugfix only (closes: bug#33152)
-- New version of DOS patch
sendmail (8.9.2-1) unstable; urgency=high, closes=30292
* New upstream source, bug fixes and a DOS fix
-- syslog filling/DOS (#30292)
sendmail (8.9.1-5) unstable; urgency=high, closes=25639
* Apply sendmail 8.9.1a security patch to protect the various MUAs
All of these could get an announcement to debian-security-announce.
> > I have also heard lot of bad things about him and his qmail, however,
> > that document is not about qmail.
>
> That very URL (having "maildisasters" and "sendmail" in it) shows me
> that he is not doing an objective analysis, so I don't see why I need
> to waste my time reading something prejudiced.
Why do I need to waste my time reading your messages, then? Please,
first read, *then* bitch, please :)
> > > And what about sendmail.mc? Is it really that hard to read? Really
> > > now, it's pretty easy.
> >
> > It is. Again, we come to the point where I say that exim.conf is *more*
> > readable. Darn. Degustibus non est disputandum.
>
> And what makes it more readable? And is this difference really
> significant? And if so, why?
In sendmail.{cf,mc}, you sometimes define variables like:
Xy<value>
Or have things like:
FEATURE(something)dnl
In exim, you would always find this form:
<short desc of Xy> = <value>
I am not an expert on public opinion etc, but it is obvious to me which
format new users would prefer.
Difference is significant since the default MTA should be one of
the most user friendly ones.
I repeat, I don't want to say anything about whether exim is more
powerful than sendmail (or anything like that), just what do I think
why exim was made default.
BTW can we stop this *pointless* thread before anyone responsible says
his reasons?
--
enJoy -*/\*- http://jagor.srce.hr/~jrodin/
Reply to: