We need a new subject- bug fixes

[Les Mikesell]

Todd Zullinger wrote:

>>> It is documented directly within the configuration file.
>> In other words you have to already know where to look before you can
>> find the change specific to the distribution.
> 
> Right, you have to know how to configure sendmail before you can begin
> to configure it.  What class of moron^wsysadmin are you advocating
> Fedora cater to in setting up a functional MTA able to send and
> receive on the internet? 

The same people who use fedora for any other task - like permitting 
inbound ssh access, which is probably more likely to allow their machine 
to be compromised.  These people are better off with standard working 
configurations than mucking around in a poorly documented config file.

> I find it to be quite wise that the daemon
> is shipped with the minimum needed settings to allow it to operate.

Perhaps you missed the point that sending is permitted by the stock 
distribution.

> By default, sendmail is needed on a Fedora box to accept mail
> generated by cron jobs and such.  It does this quite well (and hence
> isn't "broken" no matter how many times you repeat that mantra).

But none of the usual fedora invocations to activate a network service 
work.  That is the broken part.

> In order to configure sendmail to accept network connections, further
> configuration is needed and the sysadmin should definitely read the
> documentation. 

It is needed only because it isn't provided in the distribution.

> As I don't use sendmail at all (having long ago decided that it's
> configuration was made of pure crack), I don't have any interest in
> submitting patches upstream or to Fedora's sendmail maintainer(s).

So why do you think you should qualify as a cheerleader for the way the 
distribution's configuration works?  If it was handled like all the 
other programs in the distribution it would be unlikely that you would 
ever have to edit a configuration file.

> But perhaps you would be interested in doing so as you seem to be very
> convinced that this behavior should be changed or at least better
> documented.

It's kind of irrelevant to me, since I have a working configuration, but 
others seem to be interested in the popularity of the distribution and 
this is something that might affect it.  And it is certainly a 
counterexample to a policy of staying close to the defaults of the 
developer's version.

> BTW, did you ever file any bugs since the last time this was discussed
> to death?  It'd be handy to have them here for reference the next time
> this comes up.

No, I don't consider it a bug. It is clearly disfunctional as 
distributed by design.

-- 
   Les Mikesell
    lesmikesell at gmail.com