drop default MTA for Fedora 15

pbrobinson at gmail.com pbrobinson at gmail.com
Mon Aug 23 19:54:54 UTC 2010


On Mon, Aug 23, 2010 at 8:30 PM, Jon Ciesla <limb at jcomserv.net> wrote:
>  On 08/23/2010 02:21 PM, pbrobinson at gmail.com wrote:
>> On Mon, Aug 23, 2010 at 8:15 PM, Jon Masters<jonathan at jonmasters.org>  wrote:
>>> On Sun, 2010-08-22 at 20:10 +0200, drago01 wrote:
>>>> On Sun, Aug 22, 2010 at 7:45 PM, Rex Dieter<rdieter at math.unl.edu>  wrote:
>>>>> pbrobinson at gmail.com wrote:
>>>>>
>>>>>> I know its been discussed in the past but there's been reasons not to
>>>>>> drop a default MTA but now that cronie (the last actual dependency)
>>>>>> has support for logging to system logs is there any reason to include
>>>>>> an MTA by default for F-14?
>>>>> A bit late to consider for F-14 imo (I'd argue something like should in
>>>>> place and testable by or near feature freeze), F-15 is doable.
>>>> Test what? That no MTA is present?
>>>>
>>>> I'd  say we should stop arguing forever and just do it.
>>> What's the benefit of having no default MTA at all? Is it that Desktop
>>> users don't care about MTAs being installed? what about those of us who
>>> care more about server installations than Desktop?
>> In a server config I'm sure the person configuring the server would
>> know how to install a MTA, and in a lot of cases they might not want
>> sendmail but rather say postfix/exim etc. All its doing is removing it
>> from the base and core groups in comps.
>>
>> Peter
> I use it on desktop installs.  I know how to install it, but might there
> be packages that require local MTA functionality that have heretofore
> assumed it would be there, but don't Require it?  Might someone be able
> to find this out, someone with greater RPM-fu than my own?

Well the default installed MTA, which is sendmail, requires
configuration to send mail for anything other than local delivery and
all the daemons send mail to root which also requires further
configuration. Also anything that requires a MTA should require it as
per standard package guidelines.

Peter


More information about the devel mailing list