Re: [systemd-devel] [PATCH] add sprezzos as alias for debian in autoconf script
On Sun, Dec 30, 2012 at 4:09 PM, Jan Engelhardt <jengelh@inai.de> wrote:
> On Sunday 2012-12-30 15:02, Tomasz Torcz wrote:
>
>>On Sat, Dec 29, 2012 at 03:30:08PM -0500, nick black wrote:
>>> - at some point, either debian or the derivative might lurch in a new
>>> direction. in either case, carefully-crafted, minimal new
>>> TARGET_-specific code (or however you choose to do it) can mirror the
>>> divergence. all scripts continue to work.
>>
>> At this point derevative should stop and think. Is this diversion really
>>needed (hint: almost never is)? What advantages this diversion brings, if
>>any?
>
> Diversity, of course.
>
>>If it's a good idea, why upstream did not go this way?
>
> Sometimes, upstream(s) are seen as uncooperative, or simply going a
> "boring" way -
The only really boring thing is alternatives like:
/etc/HOSTNAME
/etc/hostname
/etc/sysconfig/network :: HOSTNAME=
> derivates like going where no upstream has gone
> before. There is not always just a single answer to a given problem.
> That is why there is not only sysvinit or just systemd.
What you say here does not really apply in this context. Diversity to
try something new or in a way that has the potential to develop a new
way to solve a problem is a nice thing which we surely would want to
support. But needless and trivial random legacy locations of the same
config data, which just confuses everybody is not a goal at all for
the systemd source tree/build sys support.
Kay
Reply to: