"Jóhann B. Guðmundsson"
johannbg at gmail.com
Wed Sep 14 12:16:54 UTC 2011
On 09/14/2011 11:50 AM, Rahul Sundaram wrote:
> On 09/14/2011 05:13 PM, Genes MailLists wrote
>> I realize, but that was indeed part of the point of my reply - lets
>> avoid making up things (with or without hyperbole) - and best we can,
>> stick to facts and real issues.
> You are ignoring the real issue. Since you don't seem to understand my
> point yet but let me rephrase it. Any update has a risk. A major new
> version of a init system is a substantial risk in an update. You
> shouldn't push it as an update unless there is a substantial
> justification for taking that risk. Since even a minor problem in a
> init system can result in a system that you can't easily recover from,
> the question always is not why not but why. The goal of the update
> policy is to ask that question.
Users that require newer bits that we ship in GA can just rebuild the
relevant components from srpm in koji and ofcourse they get to keep the
broken bits by doing so.
And FYI to all those that gloriously want to upgrade and claim that it's
bug free or "they" ( all of what two people ) not encountered any issues
inetd-style socket activation is borked in .35 ( users need to downgrade
to .34 or add StandardOutput=socket if they cant wait until 36 goes out ).
More information about the devel